Handbook of Cultural Heritage Analysis 3030600157, 9783030600150

Citation preview

Sebastiano D’Amico Valentina Venuti Editors

Handbook of Cultural Heritage Analysis

Handbook of Cultural Heritage Analysis

Sebastiano D’Amico • Valentina Venuti Editors

Handbook of Cultural Heritage Analysis With 1065 Figures and 126 Tables

Editors Sebastiano D’Amico Department of Geosciences University of Malta Msida, Malta

Valentina Venuti University of Messina Messina, Italy

ISBN 978-3-030-60015-0 ISBN 978-3-030-60016-7 (eBook) ISBN 978-3-030-60017-4 (print and electronic bundle) https://doi.org/10.1007/978-3-030-60016-7 © Springer Nature Switzerland AG 2022 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

The main goal of this volume is to focus on different aspects of noninvasive techniques as well as digital technologies to study and preserve cultural heritage. The handbook focuses on multidisciplinary approaches that could help site managers to better deal with their sites from conservation to prevention. The book is conceived to foster exchanges of ideas and information to update innovation on measurements suitable for cultural heritage across several disciplines. Chapters on measurements deriving from the large number of analytical methodologies and tools (spectroscopy, chemometrics, and modeling) are also of interest to the scientific community. Considering the tremendous amount of new interactive technologies, which inevitably influence traditional arts, it is important to integrate the scientific and archeological findings with the use of virtual systems to better serve museums and archaeological sites. This is not a textbook, but surely it can be used at both undergraduate and graduate levels. With the main goal of helping the reader navigate, the handbook has been organized into different parts. Each of these tackles basic scientific concepts (devoted to the theory and the science of the problem) as well as case studies to illustrate successful projects. It helps and guides the reader in comprehending the basic principles of available techniques as well as presents case studies to illustrate the application of scientific methods to cultural heritage analysis. Students and researchers from the humanities approaching scientific enquiry should find it a useful manual of available techniques and protocols, whereas scientists applying familiar techniques and methods to unfamiliar problems related to cultural heritage materials might discover interesting and unusual fields of investigation. The extensive and up-to-date reference list aims to be a useful starting point for further reading on all the presented topics. Within the handbook, we cannot claim to have provided any exhaustive panoramic of these too large thematic, but to have received contributions that express a quite updated state of art. The reader will be able to appreciate the wideness of the topic, which is multiscale and multidisciplinary, and this is a field where it is concretely possible, and indeed needed, to have a dialog between humanistic and scientific (in the sense of the so-called hard sciences) competences. We provide an overview for further studies on conservation issues, which considers both advantages and disadvantages of digital technologies in preserving cultural heritage. v

vi

Preface

Through an overview of applications and professional experiences, our ambition, with the following part on restoration and conservation, has been to highlight the new contents and methodologies adopted for diagnostics and conservation. In particular, this part illustrates innovative investigation and preservation methods for cultural heritage, touching on different, perhaps unexpected, issues that reflect its current needs. We hope that the more holistic approach to heritage in conservation programs, which has evolved in the last 20 years, may continue and spread to future generations. Its application in management plans plays a vital role in the ongoing conservation process. Msida, Malta Messina, Italy January 2022

Sebastiano D’Amico Valentina Venuti

Acknowledgments

We are grateful to all the authors for their close co-operation while preparing their chapters. We also gladly acknowledge all the referees, belonging to various research institutions located worldwide. Their careful reading and constructive suggestions contributed to the standard of the final versions of each manuscript collected in this book. Deep appreciation goes to the associated editors who professionally organized their respective parts and contributed the to the realization of this volume. Finally, special thanks go to Annett Buettner, Juby George, Johanna Schwarz, Ursula Barth, and all the editorial staff for their professional assistance and technical support during the entire publishing process. Sebastiano D’Amico Valentina Venuti

vii

Contents

Volume 1 Part I 1

.......................................

1

Introduction to Cultural Heritage Analysis . . . . . . . . . . . . . . . . . . Sebastiano D’Amico, Vincenza Crupi, Zsolt Kasztovszky, Laura Pecchioli, Raffaele Persico, Mauro Saccone, Rosarianna Zumbo, and Valentina Venuti

3

Part II

Introduction

Large Facilites and Cultural Heritage . . . . . . . . . . . . . . . . . . .

11

2

Large Facilities and Cultural Heritage Research . . . . . . . . . . . . . . Zsolt Kasztovszky

13

3

Depth-Dependent Bulk Elemental Analysis Using Negative Muons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adrian D. Hillier, Beth Hampshire, and Katsu Ishida

23

X-Ray Absorption Spectroscopy (XAS) Applied to Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Francesco D’Acapito

45

Instrumental Neutron Activation Analysis and Its Application to Cultural Heritage Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . Michael D. Glascock

69

4

5

6

7

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Zsolt Kasztovszky, Christian Stieghorst, H. Heather Chen-Mayer, Richard A. Livingston, and Richard M. Lindstrom Neutron Resonance Analysis Methods for Archaeological and Cultural Heritage Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . Peter Schillebeeckx and Hans Postma

95

145 ix

x

Contents

8

Small-Angle Neutron Scattering for Cultural Heritage Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adél Len, Katalin Bajnok, and János Füzi

9

Neutron Imaging of Cultural Heritage Objects . . . . . . . . . . . . . . . David Mannes and Eberhard H. Lehmann

10

Integration of Neutron-Based Elemental Analysis and Imaging to Characterize Complex Cultural Heritage Objects . . . . . . . . . . . . . László Szentmiklósi, Zoltán Kis, and Boglárka Maróti

11

Ancient Buddhist Metal Statues Using Neutron Tomography . . . . Eberhard H. Lehmann, David Mannes, Michael Henss, and Markus Speidel

12

Neutron Activation Autoradiography for Investigation of paintings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andrea Denker, Nikolay Kardjilov, and Birgit Schröder-Smeibidl

189 211

239 273

305

13

ECHO: The ELETTRA Cultural Heritage Office . . . . . . . . . . . . . Franco Zanini and Federico Bernardini

315

14

Study of Japanese Swords at the Neutron Source in J-PARC . . . . Yoshiaki Kiyanagi

355

15

Cultural Heritage Project at Australian Nuclear Science and Technology Organisation (ANSTO) . . . . . . . . . . . . . . . . . . . . . . . . Filomena Salvemini, Rachel White, Vladimir A. Levchenko, Andrew M. Smith, Zeljko Pastuovic, Attila Stopic, Vladimir Luzin, Mark J. Tobin, Ljiljana Puskar, Daryl Howard, Joel Davis, Maxim Avdeev, Sue Gatenby, Min-Jung Kim, Francesco Grazzi, Kenneth Sheedy, Scott R. Olsen, Carla A. Raymond, Constance Lord, Candace Richards, Joseph Bevitt, Rachel S. Popelka-Filcoff, Claire E. Lenehan, Simon Ives, Paula Dredge, Andrew Yip, Matthew Theodore Brookhouse, and Anne Gerard Austin

Part III 16

17

18

Archeometry and Portable Instruments . . . . . . . . . . . . . . .

375

443

Mass Quadrupole Spectrometry Coupled to Laser Ablation for Cultural Heritage Applications . . . . . . . . . . . . . . . . . . . . . . . . . L. Torrisi, M. Cutroneo, and A. Torrisi

445

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to Handheld Instrumentation . . . . . . . . . . . . . . . . . . . . . . Giorgio S. Senesi and Olga De Pascale

465

Provenance of Italian and Central European Archaeological Obsidians by Non-destructive WDXRF Method . . . . . . . . . . . . . . A. M. De Francesco, M. Bocci, and G. M. Crisci

505

Contents

19

20

xi

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and Characteristics for In Situ Cultural Heritage Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carlos Roberto Appoloni, Fabio Lopes, Paulo Sergio Parreira, Tiago Dutra Galvão, Fabio Luiz Melquiades, Renato Akio Ikeoka, and Eduardo Inocente Jussiani Raman Spectroscopy: Methods and Techniques for Applications in Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . Giovanna Ruello, Antonio Alessio Leonardi, Dario Morganti, Maria Josè Lo Faro, Alessia Irrera, and Barbara Fazio

519

559

21

Terahertz Waves in Archaeology I. Cacciari

..........................

581

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings . . . . . Anabelle Kriznar

613

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Anabelle Kriznar, Kilian Laclavetine, Francisco J. Ager, Claudia Caliri, Francesco Paolo Romano, and Miguel Ángel Respaldiza

24

Archaeometallurgy: A Discipline Between Past and Future Claudio Giardino

Part IV

.....

Applied Geophysics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

641

673

703

25

Ground Penetrating Radar System: Principles . . . . . . . . . . . . . . . Mezgeen Rasol, Vega Pérez-Gracia, Francisco M. Fernandes, Jorge C. Pais, Sonia Santos-Assunçao, and James S. Roberts

705

26

Integrated NDT for Building Cultural Heritage . . . . . . . . . . . . . . . Giovanni Leucci and Lara De Giorgi

739

27

System of Potential Geophysical Field Application in Archaeological Prospection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lev Eppelbaum

771

3D Electrical Resistivity Tomography for Noninvasive Characterization of Historical Walls . . . . . . . . . . . . . . . . . . . . . . . . Nasser Abu Zeid

811

Noninvasive Characterization and Monitoring of Building Elements: Electrical Resistivity Investigations . . . . . . . . . . . . . . . . Giovanni Santarato

831

28

29

xii

30

31

32

Contents

Identification of Predominant Frequencies of Selected High Rise Historical Structures in Crete . . . . . . . . . . . . . . . . . . . . . . . . . Filippos Vallianatos and Margarita Moisidi Acoustic Detection and Mapping of Submerged Stone Age Sites with Knapped Flint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ole Grøn, Lars Ole Boldreel, Rostand Tayong Boumda, Philippe Blondel, Bo Madsen, Egon Nørmark, Deborah Cvikel, Ehud Galili, and Antonio Dell’Anno Geophysics and Archaeological 3D Surface Documentation in Ostia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Axel Gering

Part V Monitoring of Cultural Heritage, 3D Survey, Models, and GIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

859

901

935

959

33

Machine Learning: A Novel Tool for Archaeology . . . . . . . . . . . . . I. Cacciari and G. F. Pocobelli

961

34

Indoor Air Quality in Heritage and Museum Buildings . . . . . . . . . 1003 Paola Fermo and Valeria Comite

35

Nanoparticles in the Field of Built Heritage Restoration: Challenges and Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1033 Silvestro Antonio Ruffolo and Mauro Francesco La Russa

36

Novel Light Cleaning Technology for Cultural Heritage Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1051 Frantisek Peterka

37

Organic Residues Analysis (ORA) in Archaeology . . . . . . . . . . . . . 1075 Silvia Polla and Andreas Springer

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens by Decay Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1121 Cristiano Riminesi, Rachele Manganelli Del Fá, Silvia Vettori, Fabio Tarani, and Piero Tiano

Volume 2 39

Mechanical Monolithic Inertial Sensors for Historical and Archeological Heritage Real-Time Broadband Monitoring . . . . . . 1137 Fabrizio Barone and Gerardo Giordano

40

Vibration-Based Procedure for the Structural Assessment of Heritage Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1167 Daniele Brigante, Carlo Rainieri, Matilde A. Notarangelo, and Giovanni Fabbrocino

Contents

xiii

41

Dynamic Investigation of Cultural Heritage Buildings for Seismic Safety Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1187 Ahmed Elyamani, Pere Roca, Oriol Caselles, and Jaime Clapes

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and Analysis Methodologies . . . . . . . . . . . . . . . . . . . . . 1221 Gianni Bartoli, Michele Betti, Luciano Galano, and Massimiliano Pieraccini

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1269 Susanna Bracci, Maria Piera Caggia, Emma Cantisani, Tommaso Ismaelli, Massimo Limoncelli, Cristiano Riminesi, Giuseppe Scardozzi, and Silvia Vettori

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1297 Laura Brambilla

45

Information and Communication Technology (ICT) for Built Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1329 Michela Cigola, Arturo Gallozzi, Silvia Gargaro, Leonardo Paris, and Rodolfo Maria Strollo

46

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage Management . . . . . . . . . . . . . . . . . . . . . . . . . . . 1351 Andrea Favretto and Bruno Callegher

47

3D Digitization of Tangible Heritage . . . . . . . . . . . . . . . . . . . . . . . 1363 George Pavlidis and Anestis Koutsoudis

48

eXtended Reality for Cultural Heritage . . . . . . . . . . . . . . . . . . . . . 1405 Ronan Gaugne, Jean-Baptiste Barreau, Flavien Lécuyer, Théophane Nicolas, Jean-Marie Normand, and Valérie Gouranton

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1439 María Mercedes Morita, Daniel Alejandro Loaiza Carvajal, and Gabriel Mario Bilmes

50

Role of Digital Technologies, Robotics, and Augmented Realities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1465 Paolo Salonia

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1497 Maurice Murphy, Stephen Fai, Lara Chow, Eimear Meegan, Simona Scandurra, Sara Pavia, Anthony Corns, and John Cahil

xiv

Contents

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural Heritage Applications . . . . . . . . . . . . . . . . . . . . . . . . . 1529 Adriana Bandiera and J-Angelo Beraldin

53

3D Metrology for Ancient Pottery Classification and Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1567 Luca Di Angelo, Paolo Di Stefano, Anna Eva Morabito, and Caterina Pane

54

Mapping Stone Age Sites by Topographical Modelling: Problems and Possibilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1595 Ole Grøn, Anton Hansson, Jessica Cook Hale, Caroline Phillips, Annabell Zander, Daniel Groß, and Björn Nilsson

55

Retrospective Photogrammetry: “Building a Time Machine” . . . . 1643 Colin Wallace

Part VI

Conservation and Restoration . . . . . . . . . . . . . . . . . . . . . . .

1677

56

Operational Modal Analysis Method for Historic Masonry Structures: Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1679 Dilek Okuyucu

57

On the Role of Historical Research in the Structural Condition Assessment of Heritage Structures . . . . . . . . . . . . . . . . . . . . . . . . . 1701 Adriana Marra, Carlo Rainieri, and Giovanni Fabbrocino

58

Conservation and Restoration Methods Applied to Ancient Ruins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1723 Laura Pecchioli

59

Sustainable Conservation and Restoration of Historical Gardens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1745 Rachele Manganelli Del Fà, Alberto Casciani, Silvia Vettori, Oana A. Cuzman, Piero Tiano, Paola Rosa, and Cristiano Riminesi

60

Geological Structural Analysis Applied to Archaeoseismology . . . 1763 Jorge Luis Giner-Robles, Miguel Ángel Rodríguez-Pascua, Raúl Pérez-López, Pablo Gabriel Silva, Teresa Bardají, Elvira Roquero, Javier Elez, and María Ángeles Perucha

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic Damage and Intensity Assessments in the Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1779 Miguel Ángel Rodríguez-Pascua, Pablo Gabriel Silva, Jorge Luis Giner-Robles, María Ángeles Perucha, Elvira Roquero, Teresa Bardají, Javier Elez, and Raúl Pérez-López

Contents

xv

62

Cleaning of Masonry Surfaces of Cultural Interest . . . . . . . . . . . . 1791 Fabio Fratini, Manuela Mattone, and Silvia Rescic

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient Pompeii . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1805 Hélène Dessales, Julien Cavero, and Agnès Tricoche

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana Tomb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1823 Ivan Roselli, Lucina Giacopini, Alessandro Colucci, Vincenzo Fioriti, Marcello Melis, Tiziana Pasciuto, Gerardo De Canio, Angelo Tatì, Francesca Boitani, and Laura D’Erme

65

The Restoration of the Acropolis of Athens: A Holistic Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1851 Maria Ioannidou, Dorina Moullou, and Dimitris Egglezos

66

Bernini’s Bust Portrait of Pope Alessandro VII Chigi at Corsini Gallery in Rome: Improving Knowledge Using Digital Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1899 Marialuisa Mongelli, Irene Bellagamba, Betarice Calosso, Giulia Chellini, Francesco Iannone, Silvio Migliori, Antonio Peroziello, Samuele Pierattini, Andrea Quintiliani, and Alessandro Cosma

67

Temporal Deformation Analysis of 3D Models as Diagnostic Tool for Panel Paintings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1915 Gianpaolo Palma, Paolo Pingi, and Eliana Siotto

68

Photogrammetric Recording of Archaeology in South-Eastern Arabia in Cultural Resource Management . . . . . . . . . . . . . . . . . . . 1933 Paul A. Yule and Michela Gaudiello

69

Italian Conservation Manuals for Pre-industrial Architecture (1989–2019) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1949 Francesco Giovanetti

70

Underground Built Heritage in Naples: From Knowledge to Monitoring and Enhancement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2001 Roberta Varriale, Mario Parise, Laura Genovese, Marco Leo, and Salvatore Valese

71

Inclusive Approach to Cultural Heritage Resilience Roko Žarnić and Barbara Vodopivec

72

Managing the Archeological Park and Open-Air Museum Viminacium (Serbia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2073 Milica Tapavički-Ilić, Emilija Nikolić, and Jelena Anđelković Grašar

73

Viminacium: Landscape and Heritage (Trans)formation . . . . . . . . 2109 Emilija Nikolić, Milica Tapavički-Ilić, and Ivana Delić-Nikolić

. . . . . . . . . . . 2037

xvi

Contents

74

Digital Documentation Management of Cultural Heritage . . . . . . 2133 Bojan Marinković, Marija Šegan Radonjić, Maja Novaković, and Zoran Ognjanović

75

Maintaining Heritage Construction Through Urban Reverse Engineering (URE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2157 Massimo Campari and Francesca D’Uffizi

76

Historical Seismic Events and Their Traces on Medieval Religious Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2181 Andrea Arrighetti, Fabio Fratini, Giovanni Minutoli, and Giovanni Pancani

77

Architects and Archaeologists in the Restoration Site . . . . . . . . . . 2211 Luigi Marino

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2225

About the Editors

Sebastiano D’Amico (Ph.D.) has been working at the University of Malta in the Department of Physics and the Department of Geosciences since 2010. He was enrolled in the physics program of the University of Messina where he was awarded the title of Dottore in Fisica; he also holds a Ph.D. in geophysics. In 2005, he moved to Rome where he joined the Istituto Nazionale di Geofisica e Vulcanologia (INGV). In 2007, he moved to the USA to join the Saint Louis University (Earth and Atmospheric Sciences Department). His research interests are in the applied aspects of geophysics and earthquake seismology. From 2016 to 2018, he served as vice president of the European Seismological Commission. In the past few years, he also worked in the field of physics applied to cultural heritage and was involved in several national and international projects. In particular, his research interests include different fields of the cultural heritage such as use and development of new technologies for the diagnostic, conservation, and restoration of cultural heritage and digitalization of monuments, statues, and paintings. Furthermore, he is involved in teaching activity on several study units offered by the University of Malta and Messina. He also taught study units in the European Union as well as the USA. He is the author of more than hundred peerreviewed publications. More details can be found at http://staff.um.edu.mt/sebastiano.damico

xvii

xviii

About the Editors

Valentina Venuti (Ph.D.) is full professor of applied physics (and also cultural heritage, environmental assets, biology, and medicine) in the Department of Mathematical and Computer Sciences, Physical Sciences and Earth Sciences, University of Messina (Italy). She is co-author of 166 publications on ISI rated journals and overall 150 communications to congresses. Citations 2633, H-index: 28. Her research interests are focused on: (1) FT-IR and Raman spectroscopies, neutron and synchrotron radiation techniques applied to cultural heritage; (2) spectroscopic study of biological and drug/carriers systems and their application in pharmaceutical field. She took part, many times as PI of the related projects selected by International User Selection Panels, in 54 experiments of neutron spectroscopy and synchrotron radiation spectroscopy carried out at European large-scale facilities. She is Responsible of the Laboratory of Applied Physics the Department of Mathematical and Computer Sciences, Physical Sciences and Earth Sciences at the University of Messina. She is a delegate of the director of the Department of Mathematical and Computer Sciences, Physical Sciences and Earth Sciences for the “Internationalization of Departmental Activities” at the University of Messina. She has been appointed, for a 3-year period 2020–2022, auditor for the Italian Association of Archaeometry. In the last 10 years, she was involved in the organization of several international conferences/workshops/schools as director, chair, program committee member, or in other positions, and in several research projects funded by MIUR and other institutions on different issues of applied physics. More details can be found at https://www.unime.it/it/ persona/valentina-venuti

Section Editors

Vincenza Crupi Dipartimento di Scienze Chimiche, Biologiche Farmaceutiche e Ambientali (CHIBIOFARAM) Università degli Studi di Messina Messina, Italy

Sebastiano D’Amico Department of Geosciences University of Malta Msida, Malta

xix

xx

Section Editors

Zsolt Kasztovszky Nuclear Analysis and Radiography Department Centre for Energy Research Eötvös Loránd Research Network Budapest, Hungary

Laura Pecchioli Institut für Archäologie Klassische Archäologie – Winckelmann-Institut Humboldt Universität Lehrbereich und OFP Berlin, Germany Brandenburgische Technische Universität Cottbus-Senftenberg BTU Cottbus Cottbus, Germany Raffaele Persico Department of Environmental Engineering University of Calabria Cosenza, Italy

Section Editors

xxi

Mauro Saccone Department of Architecture Roma Tre University Rome, Italy

Rosarianna Zumbo St Martin’s College Msida, Malta

Contributors

Nasser Abu Zeid Department of Physics and Earth Sciences, University of Ferrara, Ferrara, Italy Francisco J. Ager Centro Nacional de Aceleradores, (Universidad de SevillaCSIC-J. Andalucía), Seville, Spain Departamento de Física Aplicada I., Universidad de Sevilla, Seville, Spain Jelena Anđelković Grašar Institute of Archaeology, Belgrade, Serbia Carlos Roberto Appoloni Physics Department/CCE, State University of Londrina, Londrina, Brazil Andrea Arrighetti DSSBC – Università degli Studi di Siena, Siena, Italy Anne Gerard Austin Art Gallery of New South Wales, Sydney, NSW, Australia Maxim Avdeev Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Katalin Bajnok Neutron Spectroscopy Department, Centre for Energy Research, Budapest, Hungary Institute of Archaeological Sciences, Eötvös Loránd University, Budapest, Hungary Adriana Bandiera University of Salento, Lecce, Italy Teresa Bardají U.D. Geología, Universidad de Alcalá (UAH), Madrid, Spain Fabrizio Barone University of Salerno, Salerno, Italy Jean-Baptiste Barreau CNRS, Université Paris 1 Panthéon-Sorbonne, Paris, France Gianni Bartoli Department of Civil and Environmental Engineering (DICEA), University of Florence, Florence, Italy Irene Bellagamba Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy J-Angelo Beraldin Ottawa, Canada xxiii

xxiv

Contributors

Federico Bernardini Dipartimento di Studi Umanistici, Università Cà Foscari Venezia, Venezia, Italy Multidisciplinary Laboratory, The Abdus Salam International Centre for Theoretical Physics, Trieste, Italy Michele Betti Department of Civil and Environmental Engineering (DICEA), University of Florence, Florence, Italy Joseph Bevitt Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Gabriel Mario Bilmes Laboratorio de Ablación Láser, Fotofísica e Imágenes 3D, Centro de Investigaciones Ópticas (CONICET-CIC-UNLP), Gonnet, Provincia de Buenos Aires, Argentina Departamento de Ciencias Básicas, Facultad de Ingeniería, Universidad Nacional de La Plata, La Plata, Provincia de Buenos Aires, Argentina Philippe Blondel Department of Physics, University of Bath, Bath, UK M. Bocci Dipartimento di Biologia, Ecologia e Scienze della Terra, Università della Calabria, Rende, CS, Italy Francesca Boitani Former SBAEM, Rome, Italy Lars Ole Boldreel Department of Geosciences and Natural Resource Management, University of Copenhagen, Copenhagen K, Denmark Rostand Tayong Boumda Acoustics and Environmental HydroAcoustics Lab, Université Libre de Bruxelles (ULB), Brussels, Belgium Faculty of Creative Arts, Technologies and Science, University of Bedfordshire, University Square, Luton, UK Susanna Bracci National Research Council, Florence, Italy Laura Brambilla Haute Ecole Arc Conservation-restauration, HES-SO University of Applied Sciences and Arts Western Switzerland, Neuchâtel, Switzerland Daniele Brigante Department of Biosciences and Territory, StreGa Laboratory, University of Molise, Campobasso, Italy Matthew Theodore Brookhouse Australian National Museum, Canberra, ACT, Australia I. Cacciari Istituto di Fisica Applicata “Nello Carrara”, Consiglio Nazionale delle Ricerche, Sesto Fiorentino, Italy Maria Piera Caggia National Research Council, Lecce, Italy John Cahil Office of Public Works, Dublin, Ireland Claudia Caliri INFN, Laboratori Nazionali del Sud, Catania, Italy CNR, Istituto Scienze del Patrimonio Culturale, Catania, Italy

Contributors

xxv

Bruno Callegher Department of Humanities, University of Trieste, Trieste, Italy Betarice Calosso Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Massimo Campari Iperboole srl, Rome, Italy Emma Cantisani National Research Council, Florence, Italy Alberto Casciani Ditta Casciani Alberto Conservazione e Restauro, Florence, Italy Oriol Caselles Department of Civil and Environmental Engineering, Technical University of Catalonia, Barcelona, Spain Julien Cavero CNRS, LGP, Paris, France Giulia Chellini Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy H. Heather Chen-Mayer Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD, USA Lara Chow Carleton University, Ottawa, Canada Michela Cigola DICeM – Department of Civil and Mechanical Engineering, UNICAS – University of Cassino and Southern Lazio, Cassino (Fr), Italy Jaime Clapes Department of Civil and Environmental Engineering, Technical University of Catalonia, Barcelona, Spain Alessandro Colucci ENEA, Rome, Italy Valeria Comite Dipartimento di Chimica, Università degli Studi di Milano, Milan, Italy Anthony Corns Discovery Programme, Dublin, Ireland Alessandro Cosma Gallerie Nazionali Barberini Corsini, Rome, Italy G. M. Crisci Dipartimento di Biologia, Ecologia e Scienze della Terra, Università della Calabria, Rende, CS, Italy Vincenza Crupi University of Messina, Messina, Italy M. Cutroneo Nuclear Physics Institute, Řež, Czech Republic Oana A. Cuzman Institute of Heritage Science – CNR, National Research Council, Florence, Italy Deborah Cvikel Leon Recanati Institute for Maritime Studies, University of Haifa, Haifa, Israel Francesco D’Acapito CNR-IOM-OGG, Grenoble, France Sebastiano D’Amico Department of Geosciences, University of Malta, Msida, Malta

xxvi

Contributors

Laura D’Erme National Etruscan Museum of Villa Giulia, Rome, Italy Francesca D’Uffizi Iperboole srl, Rome, Italy Joel Davis Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Gerardo De Canio ENEA, Rome, Italy A. M. De Francesco Dipartimento di Biologia, Ecologia e Scienze della Terra, Università della Calabria, Rende, CS, Italy Lara De Giorgi National Research Council, Lecce, Italy Olga De Pascale CNR, Istituto per la Scienza e Tecnologia dei Plasmi (ISTP) – sede di Bari, Bari, Italy Ivana Delić-Nikolić Institute for Testing of Materials (IMS), Belgrade, Serbia Antonio Dell’Anno Department of Life and Environmental Sciences (DiSVA), Università Politecnica delle Marche (UNIVPM), Ancona, Italy Andrea Denker Helmholtz-Zentrum-Berlin, Berlin, Germany Beuth University for Applied Science, Berlin, Germany Hélène Dessales Ecole normale supérieure, AOROC, PSL Research University, Paris, France Luca Di Angelo Department of Industrial and Information Engineering and Economics, University of L’Aquila, L’Aquila, Italy Paolo Di Stefano Department of Industrial and Information Engineering and Economics, University of L’Aquila, L’Aquila, Italy Paula Dredge Art Gallery of New South Wales, Sydney, NSW, Australia Dimitris Egglezos Independent Researcher, Athens, Greece Javier Elez Dpto. Geología, Universidad de Salamanca, Avila, Spain Ahmed Elyamani Archaeological Conservation Department, Cairo University, Giza, Egypt Lev Eppelbaum Department of Geophysics, Faculty of Exact Sciences, Tel Aviv University, Tel Aviv, Israel Azerbaijan State Oil and Industry University, Baku, Azerbaijan Giovanni Fabbrocino Department of Biosciences and Territory, StreGa Laboratory, University of Molise, Campobasso, Italy ITC-CNR, Construction Technologies Institute, National Research Council, L’Aquila, Italy Stephen Fai Carleton University, Ottawa, Canada

Contributors

xxvii

Andrea Favretto Department of Humanities, University of Trieste, Trieste, Italy Barbara Fazio Istituto per I Processi Chimico Fisici (IPCF), Consiglio Nazionale delle Ricerche (CNR), Messina, Italy Paola Fermo Dipartimento di Chimica, Università degli Studi di Milano, Milan, Italy Francisco M. Fernandes University Lusíada – Norte, Vila Nova de Famalicão, Portugal Vincenzo Fioriti ENEA, Rome, Italy Fabio Fratini ISPC – CNR di Firenze, Florence, Italy János Füzi Neutron Spectroscopy Department, Wigner Research Centre for Physics, Budapest, Hungary Luciano Galano Department of Civil and Environmental Engineering (DICEA), University of Florence, Florence, Italy Ehud Galili The Zinman Institute of Archaeology, University of Haifa, Haifa, Israel Arturo Gallozzi DICeM – Department of Civil and Mechanical Engineering, UNICAS – University of Cassino and Southern Lazio, Cassino (Fr), Italy Tiago Dutra Galvão State University of Londrina, Londrina, Brazil Silvia Gargaro DICeM – Department of Civil and Mechanical Engineering, UNICAS – University of Cassino and Southern Lazio, Cassino (Fr), Italy Sue Gatenby Museum of Applied Arts and Sciences, Sydney, NSW, Australia Michela Gaudiello Polish Centre for Mediterranean Archaeology (PCMA), University of Warsaw, Warsaw, Poland Ronan Gaugne Univ Rennes, Inria, CNRS, IRISA, Rennes, France Laura Genovese ISPC/National Research Council of Italy, Milan, Italy Axel Gering Humboldt University, Berlin, Germany Lucina Giacopini Around Culture Srl, Rome, Italy Claudio Giardino University of Salento, Lecce, Italy Jorge Luis Giner-Robles Dpto. Geología y Geoquímica. Fac. Ciencias, Universidad Autónoma de Madrid (UAM), Madrid, Spain Gerardo Giordano University of Salerno, Salerno, Italy János Füzi: deceased.

xxviii

Contributors

Francesco Giovanetti University of Rome 3, Rome, Italy Michael D. Glascock Research Reactor Center, University of Missouri, Columbia, MO, USA Valérie Gouranton Univ Rennes, INSA Rennes, Inria, CNRS, IRISA, Rennes, France Ole Grøn Department of Geosciences and Natural Resource Management, University of Copenhagen, Copenhagen K, Denmark Francesco Grazzi Consiglio Nazionale delle Ricerche, IFAC, Sesto fiorentino, Italy Daniel Groß Centre for Baltic and Scandinavian Archaeology (ZBSA), SchleswigHolstein State Museums Foundation, Schleswig, Germany Jessica Cook Hale Department of Anthropology, University of Georgia, Athens, GA, USA Beth Hampshire ISIS Facility, STFC Rutherford Appleton Laboratory, Harwell Oxford, Didcot, UK Department of Physics, University of Warwick, Coventry, UK Anton Hansson Department of Geology, Lund University, Lund, Sweden Michael Henss Tibetan Culture Expert, Zurich, Switzerland Adrian D. Hillier ISIS Facility, STFC Rutherford Appleton Laboratory, Harwell Oxford, Didcot, UK Daryl Howard Australian Nuclear Science and Technology Organisation, Clayton, VIC, Australia Francesco Iannone Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Renato Akio Ikeoka State University of Londrina, Londrina, Brazil Maria Ioannidou Director Emerita Acropolis Restoration Service, Athens, Greece Alessia Irrera Istituto per I Processi Chimico Fisici (IPCF), Consiglio Nazionale delle Ricerche (CNR), Messina, Italy Katsu Ishida RIKEN Nishina Center, RIKEN, Wako, Saitama, Japan Tommaso Ismaelli National Research Council, Lecce, Italy Simon Ives Art Gallery of New South Wales, Sydney, NSW, Australia Eduardo Inocente Jussiani State University of Londrina, Londrina, Brazil Nikolay Kardjilov Helmholtz-Zentrum-Berlin, Berlin, Germany

Contributors

xxix

Zsolt Kasztovszky Nuclear Analysis and Radiography Department, Centre for Energy Research, Eötvös Loránd Research Network, Budapest, Hungary Min-Jung Kim Museum of Applied Arts and Sciences, Sydney, NSW, Australia Zoltán Kis Nuclear Analysis and Radiography Department, Centre for Energy Research, Budapest, Hungary Yoshiaki Kiyanagi Graduate School of Engineering, Nagoya University, Nagoya, Aichi, Japan Japan Neutron Optics Inc., Gamagori, Aichi, Japan Anestis Koutsoudis Athena-Research and Innovation Centre in Information, Communication and Knowledge Technologies, University Campus at Kimmeria, Xanthi, Greece Anabelle Kriznar Departamento de Escultura e Historia de las Artes Plásticas, Facultad de Bellas Artes, Centro Nacional de Aceleradores, Universidad de Sevilla, Seville, Spain Department of Art History, Faculty of Arts, University of Ljubljana, Ljubljana, Slovenia Mauro Francesco La Russa Dipartimento di Biologia, Ecologia e Scienze della Terra, Università della Calabria, Cosenza, Italy Kilian Laclavetine Centro Nacional de Aceleradores, (Universidad de SevillaCSIC-J. Andalucía), Seville, Spain Departamento de Física Atómica, Molecular y Nuclear, Universidad de Sevilla, Seville, Spain Centre de Recherche et de Restauration des Musées de France (C2RMF), Paris, France Centre de Recherche sur la Conservation des Collections (CRCC), Paris, France Flavien Lécuyer Univ Rennes, INSA Rennes, Inria, CNRS, IRISA, Rennes, France Eberhard H. Lehmann Laboratory for Neutron Scattering and Imaging, Paul Scherrer Institut, Villigen, Switzerland Adél Len Neutron Spectroscopy Department, Centre for Energy Research, Budapest, Hungary Claire E. Lenehan Flinders University, College of Science and Engineering, Adelaide, SA, Australia Marco Leo ISASI/National Research Council of Italy, Lecce, Italy Antonio Alessio Leonardi Istituto per I Processi Chimico Fisici (IPCF), Consiglio Nazionale delle Ricerche (CNR), Messina, Italy Dipartimento di Fisica e Astronomia, Università di Catania, Catania, Italy

xxx

Contributors

Giovanni Leucci National Research Council, Lecce, Italy Vladimir A. Levchenko Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Massimo Limoncelli University of Palermo, Palermo, Italy Richard M. Lindstrom Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD, USA Richard A. Livingston Materials Science and Engineering Department, University of Maryland, College Park, MD, USA Maria Josè Lo Faro Istituto per I Processi Chimico Fisici (IPCF), Consiglio Nazionale delle Ricerche (CNR), Messina, Italy Dipartimento di Fisica e Astronomia, Università di Catania, Catania, Italy Daniel Alejandro Loaiza Carvajal Laboratorio de Ablación Láser, Fotofísica e Imágenes 3D, Centro de Investigaciones Ópticas (CONICET-CIC-UNLP), Gonnet, Provincia de Buenos Aires, Argentina Fabio Lopes State University of Londrina, Londrina, Brazil Constance Lord University of Sydney, Nicholson Museum, Sydney, NSW, Australia Vladimir Luzin Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Bo Madsen East Jutland Museum, Randers, Denmark Rachele Manganelli Del Fà Institute of Heritage Science – CNR, National Research Council, Florence, Italy David Mannes Laboratory for Neutron Scattering and Imaging, Paul Scherrer Institute, Villigen, Switzerland Boglárka Maróti Nuclear Analysis and Radiography Department, Centre for Energy Research, Budapest, Hungary Bojan Marinković Mathematical Institute of the Serbian Academy of Sciences and Arts, Belgrade, Serbia Luigi Marino Department of Architecture, University of Florence, Florence, Italy Adriana Marra ITC-CNR, Construction Technologies Institute, National Research Council, L’Aquila, Italy Manuela Mattone Department of Architecture and Design, Polihtecnic of Turin, Turin, Italy Eimear Meegan Virtual Building Lab, Dublin, Ireland Marcello Melis Profilocolore Srl, Rome, Italy

Contributors

xxxi

Fabio Luiz Melquiades State University of Londrina, Londrina, Brazil Silvio Migliori Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Giovanni Minutoli DIDA – Università degli Studi di Firenze, Florence, Italy Margarita Moisidi Hellenic Mediterranean University, Crete, Greece Marialuisa Mongelli Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Anna Eva Morabito Department of Innovation Engineering, University of Salento, Lecce, Italy Dario Morganti Istituto per I Processi Chimico Fisici (IPCF), Consiglio Nazionale delle Ricerche (CNR), Messina, Italy Dipartimento di Fisica e Astronomia, Università di Catania, Catania, Italy María Mercedes Morita Laboratorio de Ablación Láser, Fotofísica e Imágenes 3D, Centro de Investigaciones Ópticas (CONICET-CIC-UNLP), Gonnet, Provincia de Buenos Aires, Argentina Dorina Moullou Hellenic Ministry of Culture and Sports, Athens, Greece Maurice Murphy Virtual Building Lab, Dublin, Ireland Jean-Marie Normand Ecole Centrale de Nantes, AAU UMR 1563, Nantes, France Egon Nørmark Department of Geoscience, Aarhus University, Aarhus, Denmark Théophane Nicolas Inrap, UMR 8215 Trajectoires, Rennes, France Emilija Nikolić Institute of Archaeology, Belgrade, Serbia Björn Nilsson Department of Archaeology and Ancient History, LUX, Lund University, Lund, Sweden Matilde A. Notarangelo S2X s.r.l, Campobasso, Italy Maja Novaković Mathematical Institute of the Serbian Academy of Sciences and Arts, Belgrade, Serbia Zoran Ognjanović Mathematical Institute of the Serbian Academy of Sciences and Arts, Belgrade, Serbia Dilek Okuyucu Civil Engineering Department, Erzurum Technical University, Erzurum, Turkey Scott R. Olsen Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia

xxxii

Contributors

Jorge C. Pais Department of Civil Engineering, University of Minho, Guimarães, Portugal Gianpaolo Palma National Reaserch Council, Institute of Information Science and Technologies “Alessandro Faedo”, Pisa, Italy Giovanni Pancani DIDA – Università degli Studi di Firenze, Florence, Italy Caterina Pane Department of Industrial and Information Engineering and Economics, University of L’Aquila, L’Aquila, Italy Leonardo Paris DICEA – Department of Civil, Constructional and Environmental Engineeering, Sapienza University of Rome, Rome, Italy Mario Parise University of Bari, Bari, Italy Paulo Sergio Parreira State University of Londrina, Londrina, Brazil Tiziana Pasciuto Profilocolore Srl, Rome, Italy Zeljko Pastuovic Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Sara Pavia Trinity College Dublin, Dublin, Ireland George Pavlidis Athena-Research and Innovation Centre in Information, Communication and Knowledge Technologies, University Campus at Kimmeria, Xanthi, Greece Laura Pecchioli Humboldt Universität, Winckelmann-Institut, Archäologie, Ostia Forum Projekt, Berlin, Germany

Klassische

Technische Universität Wien, Baugeschichte und Bauforschung, Berlin, Germany Vega Pérez-Gracia RMEE Department, EEBE School, Universidad Politécnica de Cataluña, Barcelona, Spain Raúl Pérez-López Instituto Geológico y Minero de España (IGME), Madrid, Spain Antonio Peroziello Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Raffaele Persico University of Calabria, Rende, Italy María Ángeles Perucha Instituto Geológico y Minero de España (IGME), Madrid, Spain Frantisek Peterka NANOTEC System s.r.o, Praha, Czech Republic Caroline Phillips Anthropology, University of Auckland, Auckland, New Zealand Massimiliano Pieraccini Department of Information Engineering (DINFO), University of Florence, Florence, Italy

Contributors

xxxiii

Samuele Pierattini Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Paolo Pingi National Reaserch Council, Institute of Information Science and Technologies “Alessandro Faedo”, Pisa, Italy G. F. Pocobelli Istituto di Scienze del Patrimonio Culturale, Consiglio Nazionale delle Ricerche, Florence, Italy Silvia Polla Freie Universität Berlin, Berlin, Germany Rachel S. Popelka-Filcoff School of Geography, Earth and Atmospheric Sciences, University of Melbourne, Melbourne, VIC, Australia Hans Postma Department of Applied Physics, RD&M, Delft University of Technology, Delft, The Netherlands Ljiljana Puskar Helmholtz-Zentrum Berlin für Materialien und Energie, Berlin, Germany Andrea Quintiliani Dipartimento Tecnologie Energetiche – Divisione ICT, ENEA, Rome, Italy Carlo Rainieri ITC-CNR, Construction Technologies Institute, National Research Council, Naples, Italy Mezgeen Rasol Laboratoire Expérimentation et Modélisation du Génie Civil Urbain (EMGCU), Université Gustave Eiffel, Champs-sur-Marne, France Carla A. Raymond Department of Earth and Environmental Sciences, Macquarie University, Sydney, NSW, Australia Silvia Rescic CNR-Institute of Heritage Science, Florence, Italy Miguel Ángel Respaldiza Centro Nacional de Aceleradores, (Universidad de Sevilla-CSIC-J. Andalucía), Seville, Spain Departamento de Física Atómica, Molecular y Nuclear, Universidad de Sevilla, Seville, Spain Candace Richards University of Sydney, Nicholson Museum, Sydney, NSW, Australia Cristiano Riminesi Institute of Heritage Science – CNR, National Research Council, Florence, Italy James S. Roberts Met Consultancy Group, Leeds, UK Pere Roca Department of Civil and Environmental Engineering, Technical University of Catalonia, Barcelona, Spain Hans Postma: deceased.

xxxiv

Contributors

Miguel Ángel Rodríguez-Pascua Instituto Geológico y Minero de España (IGME), Madrid, Spain Francesco Paolo Romano INFN, Laboratori Nazionali del Sud, Catania, Italy CNR, Istituto Scienze del Patrimonio Culturale, Catania, Italy Elvira Roquero Dpto. de Edafología, E.T.S.I. Agrónomos, Universidad Politécnica (UPM), Madrid, Spain Paola Rosa Paola Rosa Conservazione e Restauro opere d’Arte, Florence, Italy Ivan Roselli ENEA, Rome, Italy Giovanna Ruello Istituto per I Processi Chimico Fisici (IPCF), Consiglio Nazionale delle Ricerche (CNR), Messina, Italy Silvestro Antonio Ruffolo Dipartimento di Biologia, Ecologia e Scienze della Terra, Università della Calabria, Cosenza, Italy Mauro Saccone University of Roma Tre, Rome, Italy Paolo Salonia Institute of Cultural Heritage Sciences of Italian National Research Council – ISPC CNR (formerly Institute for Technologies Applied to Cultural Heritage – ITABC CNR) - ICOMOS IT (Advisor and Executive Board Member) International Committee of Architectural Photogrammetry - CIPA (Expert Member), Rome, Italy Filomena Salvemini Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Giovanni Santarato University of Ferrara, Ferrara, Italy Sonia Santos-Assunçao Department of Land Surveying and Geo-Informatics (LSGI), Hong Kong Polytechnic University, Kowloon, Hong Kong Simona Scandurra Polytechnic of Milan, Milan, Italy Giuseppe Scardozzi National Research Council, Lecce, Italy Peter Schillebeeckx European Commission, Joint Research Centre, Geel, Belgium Birgit Schröder-Smeibidl Helmholtz-Zentrum-Berlin, Berlin, Germany DIFE, Nuthetal, Germany Marija Šegan Radonjić Mathematical Institute of the Serbian Academy of Sciences and Arts, Belgrade, Serbia Giorgio S. Senesi CNR, Istituto per la Scienza e Tecnologia dei Plasmi (ISTP) – sede di Bari, Bari, Italy

Contributors

xxxv

Kenneth Sheedy Macquarie University, ACANS, Sydney, NSW, Australia Pablo Gabriel Silva Dpto. Geología, Universidad de Salamanca, Avila, Spain Eliana Siotto National Reaserch Council, Institute of Information Science and Technologies “Alessandro Faedo”, Pisa, Italy Andrew M. Smith Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Markus Speidel ETH Zürich, Zurich, Switzerland Andreas Springer Freie Universität Berlin, Berlin, Germany Christian Stieghorst Heinz Maier-Leibnitz Universität München, Garching, Germany

Zentrum

(MLZ),

Technische

Attila Stopic Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Rodolfo Maria Strollo Tor Vergata University of Rome, Rome, Italy László Szentmiklósi Nuclear Analysis and Radiography Department, Centre for Energy Research, Budapest, Hungary Milica Tapavički-Ilić Institute of Archaeology, Belgrade, Serbia Fabio Tarani National Research Council, Institute of Heritage Science, Sesto Fiorentino (Firenze), Italy Angelo Tatì ENEA, Rome, Italy Piero Tiano Institute of Heritage Science – CNR, National Research Council, Florence, Italy Mark J. Tobin Australian Nuclear Science and Technology Organisation, Clayton, VIC, Australia A. Torrisi CEDAD (CEnter of Applied Physics, DAting and Diagnostics) – Dipartimento di Matematica e Fisica “E. De Giorgi”, Università del Salento, Lecce, Italy L. Torrisi Dipartimento di Scienze Matematiche e Informatiche, Scienze Fisiche e Scienze della Terra, MIFT, Università di Messina, S. Agata(ME), Italy Agnès Tricoche CNRS, AOROC, PSL Research University, Paris, France Salvatore Valese 3D Artist – Digital Technology 3D Graphic Assisted Area, Naples, Italy

xxxvi

Contributors

Filippos Vallianatos Faculty of Geology and Geoenvironment, Department of Geophysics – Geothermics, National and Kapodistrian University of Athens, Athens, Greece Institute of Physics of the Earth’s Interior and Geohazards, UNESCO Chair on Solid Earth Physics and Geohazards Risk Reduction, Hellenic Mediterranean University Research Center, Crete, Greece Roberta Varriale ISMed/National Research Council of Italy, Naples, Italy Valentina Venuti University of Messina, Messina, Italy Silvia Vettori Institute of Heritage Science – CNR, National Research Council, Florence, Italy Barbara Vodopivec Research Centre of the Slovenian Academy of Sciences and Arts, Ljubljana, Slovenia Colin Wallace University of Waterloo, Waterloo, ON, Canada Rachel White Australian Nuclear Science and Technology Organisation, Lucas Heights, NSW, Australia Andrew Yip University of New South Wales, Sydney, NSW, Australia Paul A. Yule Language and Cultures of the Near East – Semitic Studies, Prehistory and Near Eastern Archaeology, Ruprecht-Karls-University Heidelberg, Heidelberg, Germany Annabell Zander Department of Archaeology, University of York, York, UK Franco Zanini Elettra-Sincrotrone Trieste, Trieste, Italy Roko Žarnić Faculty of Civil and Geodetic Engineering, University of Ljubljana, Ljubljana, Slovenia Rosarianna Zumbo St Martin’s College, Msida, Malta

Part I Introduction

1

Introduction to Cultural Heritage Analysis Sebastiano D’Amico, Vincenza Crupi, Zsolt Kasztovszky, Laura Pecchioli, Raffaele Persico, Mauro Saccone, Rosarianna Zumbo, and Valentina Venuti

Abstract

The new trend in scientific investigations related to the study of cultural heritage is to focus on noninvasive techniques as well as digital technologies. The handbook will focus on multidisciplinary approaches that could help site managers to better deal with their site from conservation to prevention. The book is conceived to foster exchanges of ideas and information, update innovation on measurements suitable for

S. D’Amico (*) Department of Geosciences, University of Malta, Msida, Malta e-mail: [email protected] V. Crupi · V. Venuti (*) University of Messina, Messina, Italy e-mail: [email protected]; [email protected] Zs. Kasztovszky Nuclear Analysis and Radiography Department, Centre for Energy Research, Eötvös Loránd Research Network, Budapest, Hungary e-mail: [email protected] L. Pecchioli Humboldt Universität, Winckelmann-Institut, Klassische Archäologie, Ostia Forum Projekt, Berlin, Germany Technische Universität Wien, Baugeschichte und Bauforschung, Berlin, Germany e-mail: [email protected] R. Persico University of Calabria, Rende, Italy e-mail: [email protected] M. Saccone University of Roma Tre, Rome, Italy R. Zumbo St Martin’s College, Msida, Malta e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_1

3

4

S. D’Amico et al.

cultural heritage across several disciplines. Contribution on measurements deriving from the large number of analytical methodologies and tools (spectroscopy, chemometrics, modeling, etc.) are also of interest for the scientific community. Considering the tremendous amount of new interactive technologies which inevitably impact traditional arts, a section will be focusing on to the use of virtual reality systems in museums and archaeological sites. Finally, this handbook provides a concise compilation of approved key information on methods of research, general principles, and functional relationships in well-selected fields of science and technology applied to cultural heritage. The aim of this project is to focus on aspect of noninvasive techniques (at different scales) as well as innovative approaches to study and preserve cultural heritage. The main goal is to give an overview of different techniques encouraging multidisciplinary approaches that could help site managers to better deal with their site from conservation to prevention. The handbook is conceived to foster exchanges of ideas and information, update innovation on measurements suitable for cultural heritage across several disciplines.

The investigation of precious and ancient artifacts by means of spectroscopic methods has attracted, over the last 30 years, great attention by the scientific community. This topic represents an important part of a broader discipline known as “archaeometry,” which deals with the scientific investigation of objects of historical-artistic interest both at qualitative and quantitative level. Dating, provenance, material composition, production technology, and state of conservation are some of the fundamental information, which can be obtained from the scientific investigation of findings. However, a complete knowledge of the aforementioned features requires a multidisciplinary interplay among scientists, humanists, and archaeologists, and must be always supported by an adequate comprehension of the historical and geographical background of a specific culture or a period. In this framework, the use of spectroscopic methods represents a well-established approach for the characterization, at different length scales, of the different components in cultural heritage (CH) materials. A rapidly growing number of innovative techniques, as well as many established experimental ones, are constantly being improved and optimized for the analysis of cultural heritage. Furthermore, the use of computer vision, digital models, and virtual reality represents a valuable contribution to the study of cultural heritage as well as its dissemination to the general public. Noteworthy, the uniqueness and preciousness of artworks usually need the employment of nondestructive (or at least micro-destructive) approaches, in order to keep their chemical and physical characteristics as well as their artistic and aesthetic values. In the last decades, neutron- and synchrotron radiation (SR)-based techniques have started to play a more and more important role in archaeometry, thanks to the main properties of these probes that benefit cultural heritage and archaeological studies (e.g., [2, 19]). The sensitivity, specificity, microanalysis capability, together with the possibility to carry out noninvasive and/or nondestructive investigation, justify the increasing access demand to neutron and synchrotron radiation facilities. On one side, the broad range of complex and heterogeneous materials that can be

1

Introduction to Cultural Heritage Analysis

5

investigated, such as potteries, glasses, metals, paintings, cellulosic materials, even covered by alteration layers, encompass the wide diversity encountered in archaeological sites, historical buildings, museum collections, and, more generally, in whatever context of interest for Heritage Science. On the other side, the information that these techniques can provide, concerning dating, authentication, production technology, deterioration processes and conservation methods, complementary to those obtained by conventional techniques and inaccessible by solely historical, artistic, and archaeological studies, support and facilitate transfer of knowledge in a multidisciplinary context that involves different skills, from physicists to chemists, geologists, biologists, archaeologists, restorers, and art historians. The characterization of immovable/large artifacts built into infrastructures or, for example, permanently affixed, cannot be accomplished by laboratory benchtop instrumentation. A recent innovation in this field is represented by in situ investigations by means of mobile spectrometers. Mobile and portable instruments can be easily carried onto a specific archaeological site or museum, thus avoiding all drawbacks related to transportability and micro-destructive sampling/removals. During years, a large variety of experimental techniques, such as Raman spectroscopy, X-ray fluorescence (XRF) spectroscopy, and Fourier transform infrared (FTIR) spectroscopy, originally designed and employed for laboratory-based researches, have been adapted and miniaturized for analysis in situ in cultural heritage-related fields (e.g., [5, 21]). Since the use of the first mobile instrument for archaeometry purposes, they have been successfully employed for the characterization of statues, potteries, colored tesserae, frescoes, and painted plasters, providing the possibility to perform high-quality multi-analytical investigations with no (or minimum) sample appearance alterations. Nowadays, high-performance mobile instrumentation is commercially available for researchers worldwide. In this framework, many efforts have been devoted to the optimization of the mobile instrumental performances for a fast and noninvasive in situ analysis, capable to furnish valuable information about the investigated sample with minimum damaging risk. In this sense, thanks to the ongoing improvement of instrumental components (e.g., detectors), the quality of mobile instruments is expected to increase as well, leading to the manufacturing of enhanced spectralquality mobile spectrometers having lower cost, smaller size, and advanced characteristics. Finally, even though mobile analytical equipment emerges as the last frontier in the framework of heritage investigations, in the last decades it has also been used for astrobiological, forensic, and geoscience applications, providing a new set of tools for researchers involved in interdisciplinary projects. Applied geophysics related to cultural heritage is a field continuously increasing its specific weight and relevance (e.g., [4, 6, 12]). Excavation cannot be executed in many cases, or they have to be surgically performed, because the financial resources are not sufficient. In other cases, it is not really useful to bring to light buried remains, because of the practical impossibility to maintain the integrity of the site against grave robbers and degrading. Therefore, we cannot materially preserve the whole of the testimonies of our past: we have to select what really is worth preserving. In this regards, applied geophysics is an important instrument, because

6

S. D’Amico et al.

it can produce knowledge even without excavating (e.g., [17]) or just with a reduced number of localized excavations for testing and calibrations purposes. Furthermore, it allows a selection of the places where a larger excavation is more promising. At the same time, the exigency to preserve possible damage to the testimonies of our past raises a growing need of preventive archaeology, for which applied geophysics is an essential aid. This especially holds at least up to the moment when the remains have been studied by the specialists, and we have “extracted” from them at least the meaningful pieces of knowledge that they can deliver to us. These considerations, essentially referred to archaeological sites, apply even more on historical monuments, where applied geophysics can provide details about the history and the change of use of the monument, that in many cases are not archived in documents, or maybe are incorrectly archived, or more simply the documents have been lost for any reason. This is especially challenging in urban environments, where excavations are even more difficult to be performed (the historical monument is customarily a building still in use) and interference signals (especially electromagnetic) are probably more critical. At the same time, in this case geophysical prospecting can tell us something about the condition of the monument (fractures in walls and columns, water infiltrations, presence of cavities), gathering information essential both for the restoration (and so the preservation for the future of the work of art) and the safety of the visitors [7, 8, 10, 15, 16]. As the Faro Convention 2005 told us, one of the most important engagements is to “encourage interdisciplinary research on cultural heritage, heritage communities, the environment and their inter-relationship” and the chapter on 3D survey, models, and GIS is a part of this challenge. Each paper of this section contributes to enrich knowledge about restoration, architecture, and archaeology, but the common line is represented by a workflow which started with the digitization of tangible heritage. Archaeologists, architects, engineers, computer scientists, metrologists, and archivists use 3D copies of cultural heritage to create and share knowledge. Even if the authors use a lot of different measurement techniques (range based, image based, etc.), 3D digital copies are the starting points for interpretation, restoration, documentation, monitoring, and much more (e.g., [9]). Some papers are focused on the key role of 3D digitization and give the opportunity to understand how many types of 3D survey techniques are used for Cultural Heritage and how different they are in terms of results. There are a large number of 3D digitization systems based on different methods and approaches [1, 3, 13]. Different technology solutions take into account the accuracy of digitization, the ease and speed of use, and the range of materials that can be captured or features such as reflectance, color, and geometric complexity. We can investigate the role of 3D imaging in the evolution of 3D survey techniques or we could discover some specific challenges during the acquisition of ceramic pottery. A very interesting way to use 3D survey techniques is represented by the challenge of Retrospective Photogrammetry, who use archival photographs to build photogrammetric 3D modeling of archaeological sites. The entire set of these contributions starts as a complete state of the art of 3D survey, but they can become a way to investigate the role of 3D survey and technology in the next future. 3D surveys could be more accurate, efficient, and workflow presented could simplify

1

Introduction to Cultural Heritage Analysis

7

procedures and lowering of costs, but what could we say about the relationship between heritage, knowledge, and technology? 3D survey is a starting point but one of the most important answers is Data Interpretation and Integration, and an important role should be assigned to ontology-based context models that could represent a link between models and knowledge. Exactly for this reason, Data Integration is a trend topic across this chapter, some authors use it to present their research about relational databases and archeology using GIS tools, or to enrich 3D models with semantic meanings. GIS procedure, workflow, and tools are used to link data and 3D models with BIM (Building Information Modeling) software to create HBIM (Historic Building Information Modeling) models. This kind of 3D models receive from GIS tables information about the areas, constraints of the context, relations and rules, and allow to articulate useful considerations and interpretations. Data integration between topographic modeling and GIS tools are also used to create computational models for simulating the actions and interactions of autonomous agents and find potential Stone Age settlement zones. Researchers could choose between a very large number of 3D survey techniques, but in some cases the final goal of 3D models should be to guide choices during the workflow. As the presented research shows, when the final goal is to create an extended reality experience, the survey models could be simplified, manipulated, and converted to achieve the final results. On the other hand, when the final goal is the recomposition of a piece of pottery, resolution and accuracy of the point cloud are non-derogable objectives. Citing Ernesto Nathan Rogers, who explained the typical approach of a Milanese architect in 1952, we could introduce our case studies: the famous quote “From the spoon to the town” fitting closely to our list of case studies. On one side, iron-made artifacts, here presented, are the smallest objects surveyed and this requires special workflow and special 3D survey techniques. On the other side, architectures, churches, and city blocks represent the biggest one, but these are only the extremes of the range. Among them, there are a lot of other cultural heritage objects like frescoes, bronze statues, and shipwrecks. With 3D survey and digital copy, these objects could be measured, studied, sectioned, handled, reconstructed, and transformed without damage for the original one. Using the digital copy it is possible to add data and knowledge, enhance cultural heritage, and communicate worldwide its properties to reach another aim of Faro convention: “improve access to the heritage.” Heritage conservation deals with actions and processes aimed at safeguarding the elements that define the character of a cultural resource in order to preserve its value and prolong its physical life [11, 14, 18, 20]. Preserving heritage means maintaining and increasing the asset’s value, maintaining their original shape and architectural elements of the construction, facilitating their restoration over its replacement or demolition and, when restoration is not possible, respectfully recreating the character, period, and scale. Conservation is the process of maintaining and managing change to a heritage asset by sustaining and enhancing its significance as its historical identity. In order to respect this purpose, specific measures, standards, principles, systems, actions, techniques, and intervention methods need to be continuously revised and developed in view of both technological and cultural changes, and the context to which they are applied. Thus, conservation needs to be interdisciplinary

8

S. D’Amico et al.

and touch various aspects as reflected in the dedicated section of this volume. Survey of the current status is a precondition to a meaningful intervention. Some contributions deal with natural disasters in ancient and modern times, focusing on seismic effects and related damage recognition and classification. A crucial task for a continuous updating of the recommendations/programs on risk reduction. An archaeological site which is difficult to be reached and studied due to political and physical reasons requires the time optimization and basic survey techniques. The potential difficulty should be evaluated a priori in order to choose how to document the artifacts using the right survey technology. The use of digital technologies to facilitate planning and implementing the conservation process and data dissemination related to heritage can contribute to the integrity and its overall value. Digital technologies are also beneficial for monitoring the condition of cultural assets and its environment. Combining these two aspects can help to identify degradation mechanisms and more efficient research into causes of decay, including incidentally adapted mitigation techniques to reduce those deterioration risks of cultural heritage objects. Technology is also important for some other aspects, namely to analyze historic masonry structures or evaluate the artifacts conservation state. The work on cultural heritage should not be driven just by what is possible. Critical evaluation of past and existing conservation approaches is crucial to really improve the future value of cultural heritage. Therefore, looking at the evolution of the historical development of theories and practices on conservation, the use of conservation principles in archaeological sites and on monuments ideally shown by compelling cases can be illuminating.

References 1. Adamopoulos E, Rinaudo F, Ardissono L (2021) A critical comparison of 3D digitization techniques for heritage objects. ISPRS Int J Geo Inf 10:1–10 2. Bottari C, Crisci GM, Crupi V, Ignazzitto V, La Russa MF, Majolino D, Ricca M, Rossi B, Ruffolo SA, Teixeira J, Venuti V (2016) SANS investigation of the salt-crystallization- and surface-treatment-induced degradation on limestones of historic–artistic interest. Appl Phys Mater Sci Process 122:Article number 721. https://doi.org/10.1007/s00339-016-0252-z 3. Cooper J, Wetherelt A, Zazzaro C, Eyre M (2018) From boatyard to museum: 3D laser scanning and digital modelling of the Qatar museums watercraft collection, Doha, Qatar. Int J Naut Archaeol 47(2):419–442 4. Cozzolino M, Di Giovanni E, Mauriello P, Piro S, Zamuner D (2018) Geophysical methods for cultural heritage management. Springer International Publishing, Cham 5. Crupi V, D’Amico S, Denaro L, Donato P, Majolino D, Paladini G, Persico R, Saccone M, Sansotta C, Spagnolo GV, Venuti V (2018) Mobile spectroscopy in archaeometry: some case study. J Spectrosc:Article ID 8295291. https://doi.org/10.1155/2018/8295291 6. D’Amico S, Saccone M, Persico R, Venuti V, Spagnolo G, Crupi V, Majolino D (2017) 3D survey and GPR for cultural heritage. The case study of SS. Pietro Paolo Church Casalvecchio Siculo, Kermes 107:11–15 7. D’Amico S, Colica E, Persico R, Betti M, Foti S, Barbino MP, Galone L (2020) Geophysical investigations, digital reconstruction and numerical modeling at the Batia Church in Tortorici (Messina, Sicily): preliminary results. pp 453–456 8. Daniels DJ (2004) Ground penetrating radar, 2nd edn. IEE Press, London

1

Introduction to Cultural Heritage Analysis

9

9. El-Hakim S, Picard JA, Godin G (2004) Detailed 3D reconstruction large-scale heritage sites with integrated techniques. IEEE Comput Graph Appl 24(3):21–29 10. Jol H (2009) Ground penetrating radar: theory and applications. Elsevier, Amsterdam 11. Lourenco PB (2006) Recommendations for restoration of ancient buildings and the survival of a masonry chimney. Constr Build Mater 20:239–251 12. Martinho E, Dionísio A (2014) Main geophysical techniques used for non-destructive evaluation in cultural built heritage: a review. J Geophys Eng 11(5):053001. https://doi.org/10.1088/ 1742-2132/11/5/053001 13. Mongelli M (2020) Modelli 3D e dati GIS: una loro integrazione per lo studio e la valorizzazione dei beni culturali. Archeomatica 11:18–23 14. Mongelli M, Bellagamba I, Perozziello A, Pierattini S, Migliori S, Quintiliani A, Bracco G, Tatì A, Calicchia P (2018) Photogrammetric survey to support Non Destructive Tests at St. Alexander Catacombs in Rome. In: Proceedings of the international conference of metrology for archaeology and cultural heritage (MetroArcheo2018), Cassino, 22–24 October 15. Persico R (2014) Introduction to ground penetrating radar: inverse scattering and data processing. Wiley, New York 16. Persico R, Piro S, Linford N (2018) Innovation in near-surface geophysics: instrumentation, application, and data processing methods. Elsevier, Amsterdam 17. Persico R, D’Amico S, Matera L, Colica E, De Giorgio C, Alescio A, Sammut CV, Galea P (2019) GPR investigations at St John’s Co-cathedral in Valletta. Near Surf Geophys 17:213–229. https://doi.org/10.1002/nsg.12046 18. Pitilakis K, Karafagka S, Ntinoudi O, Kalogeras I, Eleftheriou V (2018) Seismic hazard analysis of the acropolis of Athens and seismic analysis of Propylaea Colonnade. In: Proceedings of 16th European conference on earthquake engineering, pp 1–12 19. Randazzo L, Paladini G, Venuti V, Crupi V, Ott F, Montana G, Ricca M, Rovella N, La Russa MF, Majolino D (2020) Pore structure and water transfer in Pietra d’Aspra Limestone: a neutronographic study. Appl Sci 10:6745. https://doi.org/10.3390/app10196745 20. Revez MJ, Delgado Rodrigues J (2016) Incompatibility risk assessment procedure for cleaning of built heritage. J Cult Herit 18:219–228 21. Venuti V, Fazzari B, Crupi V, Majolino D, Paladini G, Morabito G, Certo G, Lamberto S, Giacobbe L (2020) In situ diagnostic analysis of the XVIII century Madonna della Lettera panel painting (Messina, Italy). Spectrochim Acta A 228:Article number 117822. https://doi.org/10. 1016/j.saa.2019.117822

Part II Large Facilites and Cultural Heritage

2

Large Facilities and Cultural Heritage Research Zsolt Kasztovszky

Contents 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Synchrotrons and Linear Accelerators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Neutron Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14 14 17 21

Abstract

Large facilities, that is, linear accelerators, synchrotrons, research reactors, and spallation sources, are multipurpose research infrastructures which have been constructed to serve many different groups of scientists and many different disciplines. Heritage Science – the study of objects of our tangible cultural heritage – is a specific discipline that also utilizes the possibilities offered by the Large Facilities, but, because of the high value of heritage objects, it prefers noninvasive investigations. In this introductory chapter, we give a brief overview of the large facilities available worldwide for Heritage Scientists. Keywords

Large facility · Photon · Neutron · Research reactor · Synchrotron · Accelerator · Spallation source

Zs. Kasztovszky (*) Nuclear Analysis and Radiography Department, Centre for Energy Research, Eötvös Loránd Research Network, Budapest, Hungary e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_2

13

14

2.1

Zs. Kasztovszky

Introduction

When investigating the physical-chemical features of various materials, we expose the target to an external effect and detect its response in a carefully designed scientific experiment. In most of the cases the “probe” to which the target is exposed is a kind of radiation (electromagnetic radiation or sub-atomic particles, i.e., electrons, protons, neutrons, or ions) and the corresponding “answer” of the target is another kind of radiation (again, electromagnetic radiation or subatomic particles). Between the impact of the “probing” radiation and the emission of the response radiation, a physical reaction on the subatomic electrons or nuclei occurs. The latter kind of reactions are called nuclear reactions. In general, these physical reactions can be denoted as X(a,b)Y, where X is the target particle – atom or atomic nucleus, Y is the “daughter” product – atom or atomic nucleus, and a and b are the incoming and outgoing radiation. In the experiments, we detect the outgoing radiation(s), in particular, γ-radiation, X-ray, UV, visible, or IR light, or emitted ions or neutrons, to obtain structural or compositional information about the target. This general scheme is valid for every experiment in pure physics, material sciences, etc. The study of Cultural Heritage objects, however, puts an extra condition on the investigations: as much as possible, any destruction of the object must be strictly avoided. At the dawn of experimental nuclear physics, as well as when archaeometric investigations started, the experiments were performed with table top arrangements, using simple radioactive sources or an X-ray tube as an irradiation source. Later on, there was an increasing demand to enhance the intensity and/or the energy selectivity of the radiation sources, to increase the spatial resolution of the scanning, and to enhance the sensitivity and selectivity of the radiation detectors used. For these purposes, scientists started to construct the so-called “Large Facilities,” typically research reactors, spallation neutron sources, linear accelerators, and synchrotrons to enhance the efficiency of the research. In most cases, these facilities are operated and supervised by national or international organizations. Their operation requires considerable human and financial resources, but in return they attract dozens of international research teams and serve hundreds of various research projects at the same time. In the last two decades, the importance of Large Facilities and also of the international cooperations in the Cultural Heritage field has significantly grown. In the following, we will try to give a brief overview of the operation of those Large Facilities where the Cultural Heritage related research is declared to be their main scientific interest. Certainly, our “inventory” cannot be considered complete, due to new facilities being built.

2.2

Synchrotrons and Linear Accelerators

A large variety of the analytical investigations can be performed using different energy electromagnetic radiations. The photons of the electromagnetic field can induce different kinds of physical reactions on the atomic electrons or even on the atomic nuclei of the target. With the help of X-ray diffraction (XRD) or inelastic

2

Large Facilities and Cultural Heritage Research

15

X-ray scattering (IXS) one can explore the structure of a sample on molecular scale. Thanks to the coherent radiation, 3D phase contrast tomography can be performed. Using X-ray fluorescence methods, one can determine the composition of a sample. The utilized electromagnetic radiation can be produced by a radioactive source, X-ray tube, or accelerator. A synchrotron is a specific type of particle accelerator in which electrons are first speed up to a few 100 MeV in a linear accelerator, and the accelerating particle beam travels in a closed-loop path. A variable magnetic field bends the particle beam into a closed path, whose radius increases with time during the acceleration process, being synchronized to the increasing kinetic energy of the accelerated particles. The synchrotron principle was invented by Vladimir Iosifovich Veksler in 1944 [1]. The accelerated electrons are then injected in a storage ring of a few hundred meter radius, in which the desired electromagnetic radiation is generated. The electromagnetic radiation generated is then used at experimental stations located on different beamlines. The facility is also called “Light Source.” The emitted electromagnetic radiation often has special features which enable the researcher to use it in cutting-edge science. The synchrotron radiation covers a broad range of the electromagnetic radiation, from infrared to the hard X-rays. The beam has high flux, high brilliance (low divergence), and high stability in time. Both circular and linearly polarized beams can be produced in pulsed mode. The most common synchrotron-based methods used in Heritage Science are the following: μ-X-ray fluorescence (μ-XRF) for elemental microanalysis; μ-X-ray absorption spectroscopy (μ-XAS) for the local chemical state determinations of selected (trace) constituents; X-ray diffraction (XRD) for the analysis of crystalline phases; SR-based FT-infrared micro-spectroscopy (SR-FT-IR) for molecular spectroscopy; X-ray computed tomography (SR-XCT) for imaging [2]. Fifty-five synchrotron light sources are reported in the database of the International Atomic Energy Agency [3]. A further nine are under design or planning. The first light source – the VEPP-3 at Siberian Synchrotron Radiation Centre (SSRC) – started its operation in Russia, in 1971, and has been modernized in 1986–1987. Its beam energy is 2 GeV [4]. Some features of the most significant light sources are summarized in Table 2.1. We have to highlight the SOLEIL synchrotron near Paris, which plays a leading role in the European Heritage Science. Its new beamline, PUMA (French abbreviation for “Photons Utilisés pour les Matériaux Anciens”), is a hard X-ray imaging beamline optimized for the scientific investigations of the heritage sciences. The 2D imaging endstation will offer a resolution of several microns with elemental (XRF), chemical (XANES), and structural (XANES and XRD) contrast. In the future, a second endstation for 3D imaging will be added [5]. Besides SOLEIL, Cultural Heritage related applications are pursued in almost all of the synchrotron laboratories. Compared to the very complicated and expensive synchrotron light sources, linear accelerators have a much simpler design. In these instruments, charged particles (electrons, protons or ions) are accelerated in a static or oscillating electric field along a linear path. The exiting beam of particles is used as a probe for the study

Stanford Synchrotron Radiation Light Source (SSRL SPEAR3) National Synchrotron Light Source (NSLS-II)

Advanced Photon Source (APS)

Brookhaven National Laboratory, New York

1.9

USA

USA

USA

3

3

7

2.4 3 4.7–5.6

Switzerland UK USA

USA

1.7 6 6.5 8 5.5

2.75

3 2.9 6

E (GeV)

Germany Germany Japan Japan Russia

France

Societe Civile Synchrotron SOLEIL, Paris Helmholtz-Zentrum Berlin DESY, Hamburg KEK-Tsukuba RIKEN, KEK-Tsukuba Budker Institute of Nuclear Physics, Novosibirsk Paul Scherrer Institute Chilton, Oxfordshire Cornell University, Ithaca, New York Lawrence Berkeley Laboratory, California Argonne National Laboratory, Argonne, Illinois SSRL, SLAC, Stanford, California

BESSY II PETRA III Photon Factory (PF-AR) SPring-8 (Super Photon Ring – 8 GeV) VEPP-4M at Siberian Synchrotron Radiation Centre (SSRC) Swiss Light Source Diamond Light Source Cornell High Energy Synchrotron Source (CHESS) Advanced Light Source (ALS)

Australia Canada France

Melbourne Saskatoon, Saskatchewan Grenoble

Australian Synchrotron Canadian Light Source European Synchrotron Radiation Facility (ESRF) SOLEIL

Country

Location

Facility name

Table 2.1 List of the most significant light sources in the world, on the basis of the inventory by the IAEA [3]

300– 500 500

100

500

400 300 200

300 100 35 100 80

500

200 500 200

I (mA)

500

476

352

500

500 500 500

500 500 508 509 181

352

500 500 352

RF (MHz)

792

234

1104

196.8

288 562 768

240 2304 377 1436 366

354

216 171 844

Circumf. (m)

Year

2015

2004

1995

1993

2001 2006 1979

1998 2009 1986 1997 1979

2006

2006 2004 1992

16 Zs. Kasztovszky

2

Large Facilities and Cultural Heritage Research

17

of various materials, or as a secondary source of probing X-rays. In the case of particle induced X-ray emission (PIXE) analysis, which is a common method to determine the elemental composition of a sample, the atomic electrons are excited by a beam of protons. The emitted characteristic X-ray radiation is detected, from which the elemental composition is determined. The method is sensitive for most of the elements from Na to U, but its disadvantage is that the protons can analyze only the outermost few tens of micrometers of the sample [6]. According to the inventory of the IAEA, more than 300 linear accelerator based laboratories are operated in the world [7]. It is worth mentioning that the Louvre Museum in Paris, France, has its own research laboratory, called AGLAE, which was established to study the art objects of the Louvre, mostly with ion beam techniques (see ▶ Chap. 56) [6].

2.3

Neutron Sources

Another subatomic particle different from electrons, photons, or protons is the neutron. The neutron, a constituent particle of the atomic nucleus, has zero electric charge. It does not exist in a free unbound form longer than 15 min, but it can be produced by spontaneous fission of 252Cf, in (α, n) nuclear reactions, in induced fission of 235U, in neutron generators, as well as in spallation reactions [8]. Since the neutrons can undergo various nuclear reactions – including elastic and inelastic scattering as well as radiative capture –, it can be used to explore the structure and composition of an unknown sample. Furthermore, unlike the electrons, protons, or ions, because of the neutron’s zero electric charge, they can penetrate the sample material to many centimeters depth. The scattering phenomenon is utilized by the methods of time of flight neutron diffraction (TOF-ND; see ▶ Chap. 52) and small angle neutron scattering (SANS; see ▶ Chap. 53). The capture-type reactions are utilized by neutron activation analysis (NAA; see ▶ Chap. 45), prompt-gamma activation analysis (PGAA; see ▶ Chap. 46), and neutron resonance capture analysis (NRCA; see ▶ Chap. 47). Even today, the most common way of producing neutrons for research is to operate a research reactor. A research reactor is a fission reactor which is built not for energy production but to be an abundant source of neutrons. According to the database of the International Atomic Energy Agency [9], more than 200 research reactors are currently functioning all over the world. Although the first research reactor was put into operation in the late 1940s (Chalk River Laboratories, Canada), and 70 are temporarily or permanently shut down, an additional 23 are still planned or under construction. In ▶ Chaps. 45, ▶ 46, ▶ 50, and ▶ 51 you will read examples of Cultural Heritage applications of neutrons pursued at the University of Missouri (MURR), the Forschungs-Neutronenquelle Heinz Maier-Leibnitz (FRM II) in Germany, the Budapest Neutron Centre, Hungary, at the NIST in the USA, at the Helmholtz-Zentrum, Germany, and at the Dingo Reactor in Australia. In contrast with the research reactor, there are some advantages when using a spallation source to produce neutrons for research purposes. In the nuclear spallation reaction, a beam of accelerated particles is used to produce a beam of neutrons. A

18

Zs. Kasztovszky

beam of protons at around 1 GeV are shot into a target made of heavy metal (Hg, Ta, or Pb). The target nuclei are strongly excited, and upon de-excitation, 20–30 neutrons are emitted per nucleus. This type of neutron production has the advantage that the beam can be pulsed with relative ease, making time of flight (TOF) techniques possible for measuring the neutron energies. In the spallation reaction the neutrons will not trigger further spallation or fission processes to produce additional neutrons. Therefore, there is no chain reaction, which makes the process noncritical. Furthermore, since there is no spent nuclear fuel produced, the amount of radioactive waste induced during the operation is much less than in the fission reactors. On the other hand, the operational costs of a spallation neutron source are much higher than those for a research reactor. At the moment, there are only a few spallation sources operating in the world (see Table 2.2), the European Spallation Source (ESS) in Lund, Sweden, being the most recent one. Though, the ESS was started in 2014, it is still under construction. It is planned to be the largest spallation source in the world, operated by the member states of the European Commission [10]. From the point of the Cultural Heritage related applications of neutron techniques, we have to call attention to the ISIS, Rutherford Appleton Laboratory in the UK, which was launched in 1985 (see ▶ Chap. 43), the SINQ, Paul Scherrer Institute in Switzerland (see ▶ Chaps. 48 and ▶ 49), and the J-PARC in Japan (see ▶ Chap. 55).

City

Los Alamos, New Mexico

Chilton, Oxfordshire

Troitsk, Russia

Villigen

Meyrin

Oak Ridge, Tennessee

Country

USA

United Kingdom

Russian Federation

Switzerland

Switzerland

USA

Oak Ridge National Laboratory, Spallation Neutron Source (SNS)

CERN: n-TOF facility

Paul Scherrer Institute, SINQ

Neutron Complex of the INR RAS

Rutherford Appleton Laboratory, ISIS

Los Alamos Neutron Science Center (LANSCE)

Name of the institution/ facility

2006

2001

1996

1993

1985

1977

Year

2200

1500

73 MeV cyclotron 590 MeV cyclotron 20 GeV proton synchrotron 940 MeV Linac

120 H+, 40 H-

200

100–125

I (μA)

870 keV C-W

800 MeV RCS 600 MeV Linac delivers 209 MeV H +, 160 MeV H

70 MeV H- Linac

665 keV RFQ

800 MeV linac

Accelerators: E, type

Table 2.2 List of the world’s spallation neutron sources, on the basis of the inventory by the IAEA [11]

1400

1400

160

80–100

P (kW)

60

50

CW

50

50

60

Freq (Hz)

700 ns

7 ns

CW

1–2 ms

120 ns

270 ns

Pulse time

Large Facilities and Cultural Heritage Research (continued)

Liquid-Hg

Pb: 0.2 Hz

RADEX: W SVZ-100: Pb Pb (D2O-cooled)

IN-06: W

Tgt Area B (UCN): W 20 Hz Tgt 4 (WNR): W 40 Hz TS-1: W 40 Hz TS-2: W 10 Hz

Tgt 1 (Lujan): W 20 Hz

Neutron Targets

2 19

Tokai-mura, Naka-gun, Ibaraki

Vancouver, BC Dongguan, Guangdong

Lund

Japan

Canada

Sweden

China

City

Country

Table 2.2 (continued)

European Spallation Source (ESS)

China Spallation Neutron Source (CSNS)

TRIUMF: UCN facility

Japan Proton Accelerator Research Complex (JPARC)

Name of the institution/ facility

2019

2018

2017

2008

Year

63

2500

1600 MeV RCS 2 GeV Linac

40

333

I (μA)

3 GeV RCS 480 MeV Hcyclotron 80 MeV H- Linac

180 MeV H- Linac

Accelerators: E, type

5000

100

5

1000

P (kW)

14

25

33

25

Freq (Hz)

2.86 ms

250,000 previous analyses of ceramics, obsidian, limestone, ochre, and other materials is a major asset for comparisons to future artifact analysis. Although the number of INAA laboratories has decreased in recent years, the capacities of the remaining laboratories have increased to keep pace with the growing demand. The future of INAA with respect to applications in archaeological science remains strong. Acknowledgments The author acknowledges his colleagues H. Neff and B.L. MacDonald for many stimulating discussions and recommendations. This chapter was written with support from the University of Missouri Research Reactor (MURR) and a grant from the National Science Foundation (#1921776).

References 1. Caley ER (1962) Early investigations. In: Caley ER (ed) Analyses of ancient glasses 1790– 1957: a comprehensive and critical survey, monographs of the Corning Museum of glass. Corning, NY, pp 13–23 2. Klaproth MH (1801) Sur quelques vitrifications antiques. In: Mémories de l’Académie Royal des Sciences et Belles-Lettres, Classe de Philosophie expérimentale, Decker, Berlin, pp 5–16 3. Davy H (1815) Some experiments and observations on the colors used in painting by the ancients. Philos Trans 105:97–124 4. Fresenius KR (1845) Chemische analyse enier Celtischen waffe. Justus Liebigs Anal Chem 53: 136–138 5. Richards TW (1895) The composition of Athenian pottery. Am Chem J 17:152–154 6. Hevesy G, Levi H (1946) The actions of neutrons on the rare each elements. Danske Vidensk Selskab Mat-Fys Medd 14:3–34 7. Finston HL, Miskel J (1955) Radiochemical separation techniques. Ann Rev Nucl Sci 5:269–296 8. Greenberg RR, Bode P, de Nadai Fernandes EA (2011) Neutron activation analysis: A primary method of measurement. Spectrochim Acta Part B 66:193–241 9. De Soete D, Gijbels R, Hoste J (eds) (1972) Neutron activation analysis. Wiley, New York 10. Ehmann WD, Vance DE (1991) Radiochemistry and nuclear methods of analysis. John Wiley and Sons, New York 11. Glascock MD (1998) Activation analysis. In: Alfassi ZB (ed) Instrumental multi-element chemical analysis. Kluwer Academic, Dordrecht, pp 93–150 12. Sayre EV, Dodson RW (1957) Neutron activation study of Mediterranean potsherds. Am J Archaeol 61:35–41

92

M. D. Glascock

13. Glascock MD, Speakman RJ, Neff H (2007) Archaeometry at the University of Missouri research reactor and the provenance of obsidian artefacts in North America. Archaeometry 49:343–357 14. Dunnell RC (1993) Why archaeologists don’t care about archaeometry. Archaeomaterials 7:161–165 15. Boulanger MT (2013) Salvage archaeometry: lessons learned from the Lawrence Berkeley laboratory archaeometry archives. SAA Archaeol Rec 13:14–19 16. Rice PM (1987) Pottery analysis: a source book. University of Chicago Press, Chicago/London 17. Orton C, Hughes M (1993) Pottery in Archaeology. Cambridge University Press, Cambridge/New York 18. Landsberger S, Yellin J (2018) Minimizing sample sizes while achieving accurate elemental concentrations in neutron activation analysis of precious pottery. J Archaeol Sci: Rep 20:622– 625 19. Blomster JP, Neff H, Glascock MD (2005) Olmec pottery production and export in ancient Mexico determined through elemental analysis. Science 307:1068–1072 20. Masucci MA, Neff H, Glascock MD, Speakman RJ (2017) The organization of ceramic production and the origins of complexity in the Late Prehispanic coastal societies of Ecuador. In: Ownby MF, Druc IC, Masucci MA (eds) Integrative approaches in ceramic petrography. University of Utah Press, Salt Lake City, pp 39–52 21. Mutin B, Minc L (2019) The formative phase of the Helmand civilization, Iran and Afghanistan: new data from compositional analysis of ceramics from Shahr-i-Sokhta, Iran. J Archaeol Sci: Rep 23:881–899 22. Bishop RL (1994) Pre-Columbian pottery: research in the Mayan region. In: Scott DA, Myers P (eds) Archaeometry of Pre-Columbian Sites and Artifacts. The Getty Conservation Institute, Los Angeles, pp 15–66 23. Bell EE, Reents-Budet D, Bishop RL, Traxler LP (2003) Early classic ceramic offerings at Copan: A comparison of the Hunal, Margarita, and Sub-Jaguar Tombs. In: Bell EE, Canuto MA, Sharer RJ (eds) Understanding early classic Copan. University of Pennsylvania Museum, Philadelphia, pp 131–158 24. Glascock MD, Braswell GE, Cobean RH (1998) A systematic approach to obsidian source characterization. In: Shackley MS (ed) Archaeological Obsidian studies: method and theory. Plenum Press, New York/London, pp 15–65 25. Glascock MD, Neff H (2003) Neutron activation analysis and provenance research in archaeology. Meas Sci Technol 14:1516–1526 26. Glascock MD, Neff H, Stryker KS, Johnson TN (1994) Sourcing archaeological obsidian by an abbreviated NAA procedure. J Radioanal Nucl Chem 180:29–35 27. Braswell GE, Glascock MD (1998) Interpreting intrasource variation in the composition of obsidian: the geoarchaeology of San Martin Jilotepeque, Guatemala. Lat Am Antiq 9:353–369 28. Glascock MD, Kunselman R, Wolfman D (1999) Intrasource chemical differentiation of obsidian in the Jemez Mountains and Taos Plateau, New Mexico. J Archaeol Sci 26:861–868 29. Ambroz JA, Glascock MD, Skinner CE (2001) Chemical differentiation of obsidian within the Glass Buttes Complex, Oregon. J Archaeol Sci 28:741–746 30. Glascock MD, Ferguson JR (2012) Report on the analysis of obsidian source samples by multiple analytical methods. Available on request from the Archaeometry Laboratory at MURR 31. Holmes LL, Harbottle G, Blanc A (1994) Compositional characterization of French limestone: a new tool for art historians. Archaeometry 36:25–39 32. Harbottle G, Holmes LL (2007) The history of the Brookhaven National Laboratory project in archaeological chemistry, and applying nuclear methods to the fine arts. Archaeometry 49:185–199 33. Coleman ME (2010) Radioanalytical multi-elemental analysis: new methodology and archaeometric applications. PhD dissertation, Department of Chemistry, University of Missouri, Columbia

5

Instrumental Neutron Activation Analysis and Its Application to Cultural. . .

93

34. Zhu J, Glascock MD, Wang C-S, Zhao X, Lu W (2012) A study of limestone from the Longmen Grottoes of Henan province, China by neutron activation analysis. J Archaeol Sci 39:2568– 2573 35. Luedtke BE (1992) Archaeologist’s guide to Chert and Flint. Institute of Archaeology, University of California, Los Angeles 36. Hoard RJ, Bozell JR, Holen SR, Glascock MD, Neff H, Elam JM (1993) Source determination of White River group silicates from two archaeological sites in the Great Plains. Am Antiq 58: 698–710 37. Malyk-Selivanova N, Ashley GM, Gal R, Glascock MD, Neff H (1998) Geologicalgeochemical approach to sourcing prehistoric chert artifacts, northwestern Alaska. Geoarchaeology 13:673–708 38. Eiselt BS, Popelka-Filcoff RS, Darling JA, Glascock MD (2011) Hematite sources and archaeological ochres from Hohokam and O’odham sites in central Arizona: an experiment in type identification and characterization. J Archaeol Sci 38:3019–3028 39. Eiselt BS, Dudgeon J, Darling JA, Paucar EN, Glascock MD, Woodson MK (2019) In-situ sourcing of hematite paints on the surface of Hohokam Red-on-Buff ceramics using laser ablation–inductively coupled plasma–mass spectrometry (LA–ICP–MS) and instrumental neutron activation analysis. Archaeometry 61:1–18 40. MacDonald BL, Hancock RGV, Cannon A, Pidruczny A (2011) Geochemical characterization of ochre from central coastal British Columbia, Canada. J Archaeol Sci 38:3620–3630 41. MacDonald BL, Hancock RGV, Cannon A, McNeill F, Reimer R, Pidruczny A (2013) Elemental analysis of ochre outcrops in southern British Columbia, Canada. Archaeometry 55: 1020–1033 42. MacDonald BL, Fox W, Dubreuil L, Beddard J, Pidruczny A (2018) Iron oxide geochemistry in the Great Lakes Region (North America): implications for ochre provenance studies. J Archaeol Sci: Rep 19:476–490 43. Popelka-Filcoff RS, Robertson JD, Glascock MD, Descantes C (2007) Trace element characterization of ochre from geological sources. J Radioanal Nucl Chem 272:17–27 44. Popelka-Filcoff RS, Miksa EJ, Robertson JD, Glascock MD, Wallace H (2008) Elemental analysis and characterization of ochre sources from Southern Arizona. J Archaeol Sci 35: 752–762 45. Gordus AA, Gordus JP (1974) Neutron activation analysis of gold impurity levels in silver coins and art objects. In: Beck CW (ed) Archaeological chemistry I. American Chemical Society, Washington, DC, pp 124–147 46. Oddy WA (1972) The analysis of gold coins – a comparison of results obtained by non-destructive methods. Archaeometry 14:109–117 47. Meyers P, van Zelst L, Sayre EV (1973) Determination of major components and trace elements in ancient silver by thermal neutron activation analysis. J Radioanal Chem 16:67–78 48. Stefanik M, Sklenka L, Huml O, Rataj J (2019) Activation analysis of Tibetan coins and thermal neutron flux measurement at the VR-1 training reactor. Radiat Phys Chem 155:304–309 49. Rapp GR Jr, Henrickson E, Allert JD (1990) Native copper sources of artifact copper in pre-Columbian North America. In: Lasca NP, Donahue J (eds) Archaeological geology of North America. Geological Society of America, Boulder, pp 479–498 50. Hancock RGV, Pavlish LA, Farquhar RM, Salloljm R, Fox WA, Wilson GC (1991) Distinguishing European trade copper and north-eastern north American native copper. Archaeometry 33:69–86 51. Hancock RGV, Fox WA, Conway T, Pavlish LA (1993) Chemical analysis of archaeological copper and brass from Northeastern Ontario. J Radioanal Nucl Chem 168:307–315 52. Anselmi LM, Latta MA, Hancock RGV (1997) Instrumental neutron activation analysis of copper and brass from the Auger site (BdGw-3), Simcoe County, Ontario. Northeast Anthropol 53:47–59

94

M. D. Glascock

53. Hancock RGV, Chafe A, Kenyon I (1994) Neutron activation analysis of sixteenth-and seventeenth-century European blue glass trade beads from the eastern Great Lakes area of North America. Archaeometry 36:253–266 54. Kenyon I, Hancock RGV, Aufreiter S (1995) Neutron activation analysis of AD 1660–1930 European copper-coloured blue glass trade beads from Ontario, Canada. Archaeometry 37: 323–337 55. Hancock RGV, Aufreiter S, Kenyon I, Latta M (1999) White glass beads from the Auger Site, southern Ontario, Canada. J Archaeol Sci 26:907–912 56. Hancock RGV, McKechnie J, Aufreiter S, Karklins K, Kapches M, Sempowski M, Moreau JF, Kenyon I (2000) Non-destructive analysis of European cobalt blue glass trade beads. J Radioanal Nucl Chem 244:567–573 57. Sempowski ML, Nohe AW, Hancock RGV, Moreau JF, Kwok F, Aufreiter S, Karklins K, Baart J, Garrad C, Kenyon I (2001) Chemical analysis of 17th-century red glass trade beads from northeastern North America and Amsterdam. Archaeometry 43:503–515

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage Zsolt Kasztovszky, Christian Stieghorst, H. Heather Chen-Mayer, Richard A. Livingston, and Richard M. Lindstrom

Contents 6.1 Principles of the Prompt-Gamma Activation Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 The Budapest PGAA Facility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Cultural Heritage-Related Applications of PGAA at the Budapest Neutron Centre (Case Studies) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1 Case Study BNC1: Provenance Research on Archaeological Stone Objects . . . . 6.3.2 Case Study BNC2: Provenance Research on Archaeological Pottery . . . . . . . . . . . 6.3.3 Case Study BNC3: Provenance Research on Historical Glass . . . . . . . . . . . . . . . . . . . 6.3.4 Case Study BNC4: Provenance Research on Archaeological Metal . . . . . . . . . . . . 6.4 The PGAA Facility at the TUM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.1 The Neutron Source and the Neutron Guide to the PGAA Instrument . . . . . . . . . . 6.4.2 Irradiation Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.3 Detectors and Spectrometers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.4 Handling the Samples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.5 Shielding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.6 Data Acquisition and Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4.7 Extensions of the PGAA Instrument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Scientific Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

96 101 104 106 109 110 111 113 113 114 114 115 115 115 117 118

Zs. Kasztovszky (*) Nuclear Analysis and Radiography Department, Centre for Energy Research, Eötvös Loránd Research Network, Budapest, Hungary e-mail: [email protected] C. Stieghorst (*) Heinz Maier-Leibnitz Zentrum (MLZ), Technische Universität München, Garching, Germany e-mail: [email protected] H. H. Chen-Mayer (*) · R. M. Lindstrom (*) Material Measurement Laboratory, National Institute of Standards and Technology, Gaithersburg, MD, USA e-mail: [email protected] R. A. Livingston (*) Materials Science and Engineering Department, University of Maryland, College Park, MD, USA e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_6

95

96

Zs. Kasztovszky et al.

6.6 Cultural Heritage-Related Applications of PGAA at the FRM II (Case Studies) . . . . . . . 6.6.1 Case Study TUM1: Chlorine in Archaeological Iron Objects . . . . . . . . . . . . . . . . . . . 6.6.2 Case Study TUM2: Archaeological Bronzes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6.3 Case Study TUM3: Provenance Analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6.4 Case Study TUM4: Rattle Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6.5 Case Study TUM5: Cosmic Impacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6.6 Further Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7 The PGAA Facility at NIST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7.1 Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7.2 Major Element Ratios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7.3 Standard-Free Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7.4 Sample Preparation and Mounting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.8 Cultural Heritage-Related Applications of PGAA at the NIST (Case Studies) . . . . . . . . . . 6.8.1 Case Study NIST1: Application of a Neutron Lens to PGAA of Oil Painting Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.8.2 Case Study NIST2: Chemically Bound Water in Smithsonian Building Stones . . . 6.8.3 Case Study NIST3: Provenance of Chinese Nephrite Jade . . . . . . . . . . . . . . . . . . . . . . 6.9 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

119 119 120 121 121 122 122 122 124 125 127 127 128 128 131 132 137 137

Abstract

Prompt-gamma activation analysis (PGAA) provides a unique opportunity to investigate a valuable material without destroying any of it. The method is perfectly applicable to investigate valuable objects belonging to our cultural heritage, as well as other materials. There are fewer than a dozen laboratories where PGAA is routinely used and only a few of them where the necessary competence to conduct cultural heritage-related research is available. We describe three laboratories, one each in Hungary, Germany, and the USA. The applicability of the method in general is independent of the physical and chemical nature of the samples. The aim of the archaeometry research is to come to conclusions about provenance, means of production, or authenticity of the objects through their compositional features. We present some characteristic examples from each of the laboratories for different kinds of archaeological objects. Keywords

Prompt-gamma activation analysis · Cultural heritage · Provenance · Stone tools · Ceramics · Glass · Metals

6.1

Principles of the Prompt-Gamma Activation Analysis

Prompt-gamma activation analysis (PGAA), also called prompt-gamma neutron activation analysis (PGNAA), is a nuclear analytical method that utilizes the radiative capture of neutrons into the atomic nuclei, i.e., the (n,γ) reaction to measure the elemental composition of an unknown sample [1]. The schematic illustration of the (n,γ) nuclear reaction, as well as the corresponding energy level scheme of the resulting excited nucleus, can be seen in Fig. 6.1. The principle of the method was

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage E-particle

neuron

A Z

X

n J4

Compound nucleus A+1 Z

X*

J1

J5

Energy Radiactive decay

Neuron capture Target

97

Promptgamma radiation

A+1 Z

J2

J3

X A Z

A+1 Z±1

X*

Decay gamma radiation

X

A+1 Z

X*

J6 A+1

A+1 Z±1

X

X Z±1

Fig. 6.1 A schematic illustration of the radiative neutron capture and the energy level diagram of the capture and the subsequent de-excitation. (Graphic: PhD Thesis of L. Szentmiklósi, CER)

known from the 1930s [2, 3], but the first applications of PGAA were performed in Saclay, Grenoble, and at the Massachusetts Institute of Technology (MIT) in the late 1960s and the early 1970s [4–8]. The PGAA method has become more widely used since the 1980s, when high-resolution gamma ray detection systems with highpurity germanium (HPGe) detectors and high-intensity neutron guides using supermirrors at research reactors became widely available [9–11]. During the analysis, we detect the characteristic γ-radiation emitted by the capturing atomic nuclei. The energies of the photons tell us which chemical elements comprise the given sample (qualitative analysis), while the intensity of the γ-radiation provides quantitative information. In the practical implementation, PGAA instruments are installed on highintensity external beams of thermal or cold neutrons, emitted by research reactors or spallation neutron sources. Although the PGAA method is based on the same nuclear reaction as the instrumental neutron activation analysis (INAA or NAA), there are some fundamental differences between them. Since in PGAA the sample is activated using an external neutron beam – contrary to NAA, where the sample is placed inside the reactor core for activation – there is no real limit to the size of the object investigated, and also it is not necessary to take sample from the object, thus damaging it. On the other hand, since, in the case of PGAA, the detection of induced γ-photons takes place at the same time as the irradiation by neutrons, almost every chemical element can be detected. Nevertheless, the sensitivities of the PGAA method for different elements vary within a wide range. The element identification is performed by the very precise determination of the energies of the emitted γ-photons [12]. A typical prompt-gamma spectrum of an obsidian sample measured at the Budapest PGAA facility, with a 9.6
107 cm2 s1 cold neutron beam, can be seen in Fig. 6.2. Since the prompt-gamma spectrum contains several hundred lines, it is essential to properly separate the peaks for the element identification. At the Budapest laboratory, we use the Hypermet-PC software to fit the peaks in the prompt and decay spectra [13–15]. The key parameters (energy, relative intensity, partial gamma ray production cross section) of the most intense γ-photons emitted by every chemical element after neutron capture have been measured in standardization measurements and are compiled in our PGAA library.

98

Zs. Kasztovszky et al.

0

1000

2000

3000

4000

5000

1000

B 477.6 keV

Si 1273.4 keV Si 2092.9 keV

100

Si 3539 keV Si 3660.7 keV

Si 4933.8 keV 1

H 2223.3 keV

10

0.1

1

Count rate (cps)

6000 10

0.01

Si 6379.7 keV Si 7199 keV 0.1

0.001

Fe 7631.1 keV Fe 7645.5 keV 0.01

Si 8472 keV

0.0001

0.001

0.00001

0.0001

0.000001

0.00001 6000

7000

8000

9000

10000

11000

0.0000001 12000

Energy (keV)

Fig. 6.2 A typical PGAA spectrum of an archaeological stone made of gray flint. The measurement was done at the cold neutron beam of the Budapest PGAA facility. The acquisition time was 7200 s

The quantitative analysis is based on the following equation: AE ¼ m∙


 NA ∙Θ∙σ 0 ∙I γ ∙Φ0 ∙e Eγ ∙t M

ð6:1Þ

where AE is the peak area of a given element determined from the spectrum; m is the mass of a given element; NA is the Avogadro number, divided by the M molar mass of the element; Θ is the isotopic abundance; σ 0 is the neutron capture cross section; and Iγ is the gamma yield for a given nuclear reaction – all of them being nuclear constants. Φ0 is the intensity of the neutron beam and ε(Eγ) is the efficiency of the HPGe detector at a given gamma ray energy, Eγ. These latter two quantities are characteristics of the actual experimental setup. The energy dependence of the detector efficiency must be checked regularly by calibration measurements, using radioactive source and (n,γ) reactions [16]. The ε(Eγ) efficiency function of the Budapest PGAA detector can be seen in Fig. 6.3. Equation 6.1 can be written as AE ¼ m∙S∙t

ð6:2Þ

from which one can see that the sensitivity S depends partly on nuclear parameters (M, Θ, σ 0, and Iγ) and also on parameters related to the experimental setup (Φ0 and ε(Eγ)) [1].

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

99

Fig. 6.3 Efficiency curve of the gamma detectors of the Budapest PGAA facility, measured with several radioactive calibration sources and (n,γ) reactions and produced by the Hypermet-PC software. The energy shown is from 50 keV to 11 MeV

The sensitivities of the PGAA method for different chemical elements are primarily determined by the neutron capture cross sections of the atomic nuclei, which is a nuclear property of each given nucleus. The higher the neutron capture cross section, the higher the sensitivity (or equivalently the lower the detection limit) with which a given element can be measured using the PGAA method. It is also well known in nuclear physics that for low neutron energies, the neutron capture cross section, i.e., the probability of the neutron capture, is inversely proportional to the square root of the neutron energy (1=v – law). This implies that by using low-energy – in other words cold – neutrons for PGAA, one can significantly increase the sensitivity of the method. In the case of the Budapest PGAA system, thanks to the installation of the so-called cold neutron source (CNS) in 2000, the sensitivity has increased by a factor of 20 which has resulted in three possibilities: (i) the possibility of shorter measurements to determine the same components, (ii) the possibility to measure smaller samples in the same measurement time, or (iii) to detect components which were not possible to detect using thermal neutrons with the same measurement time. The elements with the highest neutron absorption cross section, such as B, Cd, Hg, Nd, Sm, Gd, and Eu, can be detected with the highest sensitivities, even in concentrations less than 1 ppm. Other chemical elements can be detected from the

100 1

Zs. Kasztovszky et al. H

Z

„ „ „ „ „ †

0.3326 b 82.02 b

3

Li 4 70.5 b 1.37 b

11

19

Rb 38

0.38 b 6.8 b

Sc 22

27.5 b 23.5 b

Y 40

1.28 b 7.70 b

Ba 57

1.1 b 3.38 b

Ti 23

6.09 b 4.35 b

La 72

8.97 b 9.66 b

12.8 b 13 b

V 24

5.08 b 5.10 b

Zr 41

0.185 b 6.46 b

Cr 25 Mn 26

3.05 b 3.49 b

13.3 b 2.15b

2.48 b 5.71 b

Ta 74

20.6 b 6.01 b

105

20 b 6.3 b

W 75

18.3 b 4.60 b

Fe 27

2.56 b 11.62 b

Nb 42 Mo 43 (Tc) 44

1.15 b 6.255 b

Hf 73

104.1 b 10.2 b

87 (Fr) 88 (Ra) 89 (Ac) 104

-

767 b 5.24 b

Re 76

89.7 b 11.5 b

Ru 45 2.56 b 6.6 b

Co 28

37.18 b 5.6 b

Rh 46

144.8 b 4.6 b

Os 77

16.0 b 14.7 b

Zn 31 2.75 b 6.38 b

Ag 48

63.3 b 4.99 b

Pt 79

10.3 b 11.71 b

Cu 30 3.78 b 8.03 b

Pd 47 6.8 b 4.48 b

Ir 78

425 b 14 b

Ni 29

4.49 b 18.5 b

Au 80

98.65 b 7.73 b

Hg 81

1.9 b 11.51 b

F 10 0.0096 b 4.018 b

S 17

0.53 b 1.026 b

As 34

33.5 b 16.8 b

0.039 b 2.628 b

Ar

0.675 b 0.683 b

Br 36

6.9 b 5.90 b

Te 53

4.7 b 4.32 b

Ne

Cl 18

Se 35

11.7 b 8.30 b

Sb 52

4.91 b 3.90 b

Pb 83

0.171 b 11.12 b

0.00019 b 4.232 b

4.5 b 5.50 b

Sn 51

0.626 b 4.892 b

O9

P 16

0.172 b 3.312 b

Ge 33

2.20 b 8.60 b

Tl 82 3.43 b 9.89 b

N8

Si 15

0.171 b 2.167 b

In 50

193.8 b 2.62 b

372.3 b 26.8 b

0.00350 b 5.551 b

Ga 32 2.75 b 6.83 b

Cd 49

2520 b 6.5 b

C7

Al 14 0.231 b 1.503 b

Sr 39

1.28 b 6.25 b

Cs 56

29.0 b 3.90 b

B6

13

Ca 21

0.43 b 2.83 b

He 0.007 b 1.34 b

5

0.063 b 3.71 b

K 20

2.1 b 1.96 b

55

V - capture V - scattering

< 10 ng < 1 Pg < 10 Pg 100 Pg no data

Na 12 Mg

0.530 b 3.28 b

37

Be

0.0076 b 7.63 b

2

Detection Limit

Element

Kr

25 b 7.68 b

I 54

6.15 b 3.81 b

Xe

23.9 b -

Bi 84 (Po) 85 (At) 86 (Rn)

0.0338 b 9.156 b

-

-

-

106

-

58

Ce 59

0.63 b 2.94 b

90

Pr 60

11.5 b 2.66 b

Nd 61 (Pm) 62 Sm 63

51 b 16.6 b

Th 91 (Pa) 92

7.37 b 13.36 b

200.6 b 10.5 b

7.57 b 8.9 b

168.4 b 21.3 b

5922 b 39 b

Eu 64 Gd 65

4530 b 9.2 b

49700 b 180 b

Tb 66

23.4 b 6.84 b

Dy 67

994 b 90.3 b

Ho 68

64.7 b 8.42 b

Er 69 Tm 70

159 b 8.7 b

100 b 6.38 b

Yb 71

34.8 b 23.4 b

Lu

74 b 7.2 b

U 93 (Np) 94 (Pu) 95 (Am) 96(Cm) 97 (Bk) 98 (Cf) 99 (Es) 100 (Fm) 101(Md)102(No) 103 (Lr) 175.9 b 14.5 b

1017.3 b 7.7 b

Fig. 6.4 The approximate detection limits of the Budapest PGAA system. Different colors indicate different ranges of sensitivities. (Graphic: Zs. Révay, CER)

concentration ranges of 103–101 weight%. The elements which can be detected with the lowest sensitivities are C, N, O, F, Sn, Pb, and Bi, because of the low neutron capture cross sections [17]. The approximate sensitivities of the Budapest PGAA instrument for the chemical elements can be seen in Fig. 6.4, where the different ranges of sensitivities are indicated by different colors. From among the most easily detectable elements, B and the rare-earth elements (Nd, Sm, Gd, and Eu) are always present in samples of geological origin, and thus they have great importance in provenance studies of some cultural heritage-related material [18]. It is also worth mentioning that H and Cl are also extremely easy to measure elements using PGAA, in concentrations of some 10 ppm, and they can provide information on the state of deterioration of many cultural heritage objects due to environmental effects (weathering, corrosion, etc.) [19]. From Eq. 6.1, one could directly calculate the amount of a given element (m) from the peak areas (AE) determined from the measured spectra. However, introducing the so-called k0,C constants, which can be calculated for every gamma line of every element, as follows:
k0,C ðXÞ ¼


θ  σ 0  I γ =M

θ  σ 0  I γ =M

 X C

one can calculate the mass ratios of two arbitrary elements using

ð6:3Þ

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

mX AX Sγ,Y AX k0,C ðY Þ eγ,Y ¼  ¼   mY AY Sγ,X AY k0,C ðXÞ eγ,X

101

ð6:4Þ

where X and Y denote two arbitrary elements and C denotes the comparator element. In most of the PGAA laboratories, H or Cl is used for comparator elements, and the k0-factors are measured in comparator measurements and tabulated in the PGAA library. By using the comparator method, one can avoid having to make absolute measurements of many of the quantities found in Eq. 6.1; only relative measurements will suffice, since their absolute calibrations will cancel out. P Supposing that all the major constituents are detected using the k0-method, i.e., mi m ¼ 1, then the concentration of the components can be calculated from the mass ratios and can be expressed as atomic%, mass%, weight%, or ppm. Since oxygen is one of the least detectable elements and, in most cases, geological samples occur as oxides, one can calculate the corresponding amount of oxides and of oxygen from the elemental concentrations, using the oxidation number. The above detailed calculations are performed using the ProSpeRo software at the Budapest PGAA laboratory [20]. There are about 20 research centers in the world where there exists the possibility, in principle, to perform prompt-gamma activation analysis, but only a few of them use the method in their everyday research program [21]. The PGAA laboratories currently operating can be found in the USA (NIST – Maryland, Texas), in Germany (FRMII – Garching), and in Hungary (BNC – Budapest). Among these, the Budapest laboratory – installed at the Budapest Research Reactor – has a 20-year-long tradition in the application of PGAA and other nuclear techniques in the research of our cultural heritage.

6.2

The Budapest PGAA Facility

The Budapest Research Reactor was constructed 60 years ago, in 1959, on the campus of the former Central Research Institute of Physics of the Hungarian Academy of Sciences, in Budapest, Hungary. The water-cooled, water-moderated tank-type (VVR) reactor originally was operated with 2 MW power. Following two comprehensive reconstructions, it started again in 1993 with 10 MW power. Starting in 1993, new external neutron beam guides have been constructed to provide neutrons for new research instruments. Three main horizontal guides lead the neutrons from the reactor core to the experimental instruments in the “Cold Neutron Source Hall” (see Fig. 6.5). The neutron guides are hollow tubes made of glass plates and coated with a neutron reflecting Ni-multilayer, called “supermirror.” The role of the supermirror guides is to conduct the neutron beams to the instruments without significant loss of their intensity. In addition to neutron reflectometry, in-beam Mössbauer spectroscopy, and smallangle neutron scattering (SANS) – which are basically meant to perform structural

102

Zs. Kasztovszky et al.

PSD

TAS

CNS Measuring Hall Shielding tunnel (for 3 neutron guides)

MTEST

BIO

REFL

NG-1

JBNS

NG-3

TASC

SANS

PGA NIPS

NG-2

Shutters (3 pcs.) SNR

Reactor Hall

Entrance to the reactor hall

DNR

TOF – Time of flight spectrometer (under construction) DNR – Dunamic neutron radiography SNR – Static neutron radiography BIO – Port used for biological experiment MTEST – Material testing diffractometer TAS – Triple axis spectrometer

TOF (neutron guide)

Cold neutron instruments: REFL – Reflectometer TASC - TAS – Triple axis spectrometer on CNS

TOF (measuring hall)

SANS – Small-angle scattering spectrometer PGAA – Prompt gamma activation analysis NIPS – Neutron-induced prompt gamma-ray spectrometer IBMS – In-beam Mössbauer spectrometer (under construction)

PSD – Powder diffractometer

Fig. 6.5 The floor plan of the Budapest Reactor Hall and the connected CNS Measuring Hall. (Graphic: F. Gajdos, CER)

studies on various materials – the PGAA and NIPS-NORMA instruments can be used to measure the elemental composition of samples with various sizes and physical-chemical forms. The first steps of the construction of the Budapest PGAA facility were made in 1995, after installation of the neutron guides in the “Cold Neutron Source Hall.” The PGAA facility is located near the end position of the neutron guide No. 10/1. The first test measurements were done in 1996. In the following years, the standardization of the method, the gradual decrease of the spectral background, the improvement of the signal-to-noise ratio, and the sample environment were all persistently worked on. In 2000, a so-called cold neutron source was installed next to the reactor core to provide low-energy (16 K equivalent temperature) neutrons for some experiments. The use of cold neutrons enhances the sensitivity of the PGAA method by a factor of 20 (see below for details), compared to that of the same intensity thermal beam. In the following, we describe briefly the present state of the Budapest PGAA facility [22]. Before the main beam in our guide enters the PGAA experimental area, it is divided into two sub-beams (upper and lower) by suitable collimators; the upper beam serves the PGAA facility and the lower beam the NIPS-NORMA setup. The thermal equivalent neutron flux at the PGAA sample positions is 9.6
107 cm2 s1. The beam can be collimated to a maximum cross section of 20
20 mm2. The intensity of the incoming neutrons is monitored and recorded with an ORDELA

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

103

Model 4511 N neutron detector throughout the whole reactor campaign of a given measurement. For special experiments, a pulsed beam can also be used. Modulation in the order of milliseconds can be done by a revolving chopper blade, while longer on-off periods can be achieved with a fast beam shutter. The experimental area is a 3
6.5 m2 space at the rear end of the “Cold Neutron Source Hall” (see Fig. 6.5). The neutrons enter the cabin and fly through a 3-m-long evacuated aluminum flight tube across the experimental area, to the beam stop placed at the back wall of the guide hall. A pneumatically actuated instrument shutter is used to control the entry of the neutron beam into the cabin, while two computer-controlled secondary shutters are in place to allow independent operation of the PGAA and NIPS-NORMA facilities. Independent sections of the modular aluminum flight tube can easily be removed and reinstalled as needed. A series of appropriate size collimators are used for each beam. The PGAA station is typically used to measure smaller pieces – up to 10 cm in height and 5 cm in diameter, while the NIPS-NORMA setup is used to analyze larger objects [22]. The PGAA target chamber is 1.5 m from the end of the guide. The sample chamber can be evacuated or filled with gases to decrease beam-induced background. To prevent scattering of neutrons into the PGAA sample from the lower beam, a layer of neutron absorber is placed below the sample. The samples are supported on thin Al frames by Teflon strings. Optionally, an automated sample changer with a capacity of 16 samples can be used. A neutron absorber after the PGAA target chamber stops the upper beam. The gamma detector system of the PGAA facility consists of an n-type HPGe (Canberra HPGe 2720/S) main detector with closed-end coaxial geometry and a BGO Compton suppressor surrounded by a 10-cm-thick lead shielding (see Fig. 6.6). The sample-to-detector distance is adjustable, but it is typically 230 mm. By removing the front detector shielding, the HPGe main detector can be placed as close as 120 mm from the target [23]. Fig. 6.6 HPGe detector (in the center), which will be surrounded by the bismuth germanate (BGO) detector, during assembling of the Budapest NIPS-NORMA experimental station. The lead shielding is not shown here. (Photo: L. Szentmiklósi, CER)

104

Zs. Kasztovszky et al.

The BGO annulus and catchers around the HPGe detect most of the photons which Compton scatter out of the Ge detector, hence depositing less energy in the detector. If the events from the HPGe and the BGO are collected in anticoincidence mode, Compton-suppressed spectra can be acquired, e.g., most of the Compton events are not in the spectrum, thus reducing the effective background. An analogue spectroscopy amplifier combined with an ADC and an Ethernet-based multichannel analyzer (Canberra AIM 556A) collects the spectral counts. Another option for recording the data is to use a CAEN N6724 digitizer card with a four-channel input and fiber optic link for list mode data acquisition. The PGAA method can be used to analyze solid, liquid, or even gaseous samples. In most practical cases, we analyze solid materials. In heritage science measurements, we can investigate samples of geological origin, i.e., various rocks, soils, clay, ceramics, glass, ores, metal, or composite samples, including any combinations of the above, sometimes including even organic material. During the measurements, up to a few grams of the sample is packed into a thin fluorinated ethylene propylene (FEP) film and placed in the neutron beam. The prompt-gamma spectra are acquired during the irradiation with neutrons. A typical acquisition takes a time varying from a few minutes up to 12–24 h, depending on the amount of sample material and the particular chemical elements one wishes to quantify. Following the irradiation, some minor amount of induced radioactivity may be produced, but, thanks to the low intensity of the neutron beam, it decays within a few hours or at maximum in some days. After storage for the “cooling” period in a safe place, the object can be returned to the curator, can be exhibited, or can undergo further investigations. It is checked in specific experiments that no visible or invisible damage to the object is caused by PGAA.

6.3

Cultural Heritage-Related Applications of PGAA at the Budapest Neutron Centre (Case Studies)

Since neutrons have zero electric charge, they do not interact with the atoms of the target, by Coulomb forces which are much stronger than the nuclear interaction. This implies that neutrons easily penetrate the sample and hence are ideal to probe the inner part of an unknown sample. If we utilize the (n, γ) reaction to determine the elemental composition, γ-photons with their high penetration properties can provide average compositional information for the whole irradiated volume. Furthermore, thanks to the relatively low-intensity (106–109 cm2s1) external neutron beams, no visible or invisible transformation of the sample material can be observed, and the induced radioactivity decays within a few days. In addition, the external neutron beams used offer the possibility to analyze certain selected parts of large objects. It follows from all these considerations that prompt-gamma neutron activation analysis using external thermal or cold beams can be considered as a completely nondestructive method that is uniquely and perfectly applicable to study valuable objects, including objects of our cultural heritage. The scientists at the Budapest

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

105

PGAA laboratory have recognized this advantage and potential as early as the end of the 1990s, right after launching the PGAA experiments at the Budapest Neutron Centre [24, 25]. In cooperation with Hungarian archaeologists and other cultural heritage experts, we have systematically explored what kinds of questions about which types of materials from the various Humanistic discipline might be answered using PGAA. In the following, we will present some case studies from our work, where PGAA proved to be a successful tool. In our research portfolio, analysis of archaeological stone materials, ceramics, glass, and metals can be considered as successful applications. However, none of the existing analytical methods is completely successful. We will show the advantages but also the limitations of the PGAA method. There are several types of questions that can be answered with the help of analytical methods. The most important questions are associated with the age of an object, with the origin of the raw materials of different kinds of objects, with the identification of a certain production technique or workshop, and with the determination of the present condition of an object. The discipline used to be called archaeometry, but more recently it is called heritage science. Dating of objects is usually performed using a few dedicated methods, including the famous C-14 dating [26, 27], the less well-known thermoluminescence (TL) [28], or optically stimulated luminescent (OSL) dating and dendrochronology [29]. In most cases, PGAA cannot do anything in direct dating, although one can indirectly obtain chronological information through the identification of a workshop or production technique. The compositional information can be utilized best in the provenance studies, i.e., when we determine the origin of a specific raw material of various objects found with or without archaeological context. The association of the object with potential sources at certain locations can assist in the work of the archaeologists, i.e., in reconstruction of the ancient trade routes, social connections, and movements of communities or even of individuals. In the history of mankind, different types of raw materials have become available and have been popular for a given time period to use to produce everyday tools and other artifacts. During the Paleolithic era (500,000 to 30,000 B.P. in Europe), the relatively easily available rocks were utilized. The shape of the tools has been formed by chipping techniques. Later, in the Neolithic period (10,000 to 6000 B.P. in Europe), the elaboration of polished stone tools became much more sophisticated. Fortunately, thanks to the exclusively physical processing, the original chemical composition of the rocks was not changed. This allowed the scientists to trace the geological sources of their raw material, based on their composition or, as it is also called, based on their chemical “fingerprinting” [30]. Already in the Neolithic, the use of another type of material for making everyday utensils, e.g., ceramics, started to spread from one community to another. For the Neolithic era, fragments of pottery have become the most abundant archaeological finds. Unlike stone objects, the composition of pottery material is determined by the craftsmen. A potter uses clay and temper to form the vessel or other object and sometimes puts glaze on the surface for decorative purposes. It follows that the material of ceramics is far from homogeneous and its composition cannot be easily

106

Zs. Kasztovszky et al.

associated with one raw material source. Instead, provenance research for pottery requires a combination of complementary analytical techniques (including microscopy, petrography, determination of the bulk major and trace element composition), in order to identify “fingerprint-like” characteristics. After the Neolithic, with the discovery of metal production (Copper Age – Bronze Age – Iron Age between 6000 B.P. and 3000 B.P.), the provenance research of metallic products has become a much more complex task. On the one hand, the origin of metallic ores in principle can be determined, based on fingerprinting geochemical elements, but in most cases it requires very sensitive analytical techniques to determine trace elements or even isotopic composition. On the other hand, in later historical periods, especially from the Bronze Age, deliberate variation of the alloying components (i.e., Cu-Sn-Pb-As in bronze, Cu-Zn in brass, or Ag-Cu in silver alloys) as well as recycling of old metal objects has become commonplace. All these circumstances make the provenance of metals a complex analytical task [31]. The situation is very similar with glass. The production of glass is claimed to have started in Mesopotamia or in Egypt around 4000 B.P. The basic constituents of glass are the SiO2 (sand), Na2O (soda), and CaO (lime), as well as some minor additives (Co, Cu, Mn or Fe) for coloring or decoloring. MgO and Al2O3 were also added. Na2O was replaced by K2O in continental regions, where wood ash was available, instead of sea-plant ash. Although in the beginning the constituents were added randomly, later well-defined recipes have been used. The amount of the basic constituents and also some contaminants at trace level might be characteristic for a certain period or production center, but the phenomenon of recycling again makes the provenance research more complicated [32].

6.3.1

Case Study BNC1: Provenance Research on Archaeological Stone Objects

As we wrote above, the chemical and mineralogical composition of lithic material is not substantially affected during the production process in prehistory nor in the “afterlife” of the object. Thus, provenance research of prehistoric lithic objects can have easy success. The key point in the research, however, is to find fingerprint-like patterns of the geochemical components at major, minor, or trace level. Certainly, the success of the research depends on the investigated material and the methods used. In order to obtain reliable results, a statistically representative series of analyses are carried out using – as much as possible – fast and low-cost methods. On the other hand, the analytical result must be precise enough and representative for the investigated material. The application of complementary methods is recommended whenever it is possible. Finally, efforts must be made to preserve the studied samples for possible further studies or to exhibit them for the public. A large portion of the prehistoric artifacts are made of obsidian – a glassy volcanic material – and most of our analytical investigations is made on that material. Fortunately, due to the specific formation process, there exist significant geochemical fingerprints for the main obsidian sources in the world, and thus one

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

107

can easily classify the sources, based on their chemical composition. In order to identify and quantify the key fingerprinting elements, several methods – both destructive and nondestructive, such as NAA [33], XRF [34], ICP-MS [35], and PGAA [36] – can be used. PGAA enables one to measure the major geochemical components, except MgO, which is typically below the quantification limit. According to our research, B, Cl, and Ti are the best discriminative which elements can be easily measured by PGAA [37]. We started to investigate the provenance of obsidian artifacts at the Budapest PGAA laboratory in 2003. Since then, we have analyzed more than 200 archaeological pieces and around 150 geological reference samples. We have focused our attention on the distribution of archaeological obsidian artifacts in the Carpathian Basin and its surroundings. We have analyzed pieces from Bosnia-Herzegovina [38], Croatia [38], Romania [39], Serbia, and Poland [40]. When we tried to compare the composition of the archaeological samples with those of the representative geological reference samples, we tried to include all the archaeologically significant geological sources, including the Carpathian sources (C1, C2E, C2T, and C3 subtypes), as well as from Lipari, Sardinia, Melos, Antiparos, Pantelleria, Palmarola, Armenia, and Anatolia. Using PGAA, we were able to differentiate macroscopically similar but chemically very dissimilar material from obsidian with other compositions, even in the beginning of the research [41]. Furthermore, all the identified archaeological obsidian pieces could be assigned to one of the main raw material types. It was a remarkable success of the method, when we were able to define a border zone between two distribution areas for the Carpathian and the Lipari obsidians in the Dinaric Mountains (Fig. 6.7) [38]. In another case study, we have proved the appearance of the good-quality Carpathian 1 obsidian at approx. 700 km northwest from its source, in the Polish Lowlands [40]. Another large group of prehistoric raw materials, called silex, comprises flint, radiolarite, chert, and jasper, which are siliceous rocks – typically having a highsilica-content composition and various conditions of formation: shallow-water sedimentary rocks (chert and flint), bathyal (deep sea/oceanic) sediments (radiolarite), and hydrothermal/limnic origin (jasper and limnoquartzite). It is important to note from the analytical point of view that despite their different colors and textures, they all contain over 95 wt% SiO2. Even so, some slight differences can be shown between the subcategories, by applying the principal component analysis on the measured compositional data [42]. Taking into consideration the detection limits of PGAA, only a few chemical components can be used for discrimination between various sources of silex-type materials. In a particular study of the chipped stone archaeological material from the Erősd (Arius‚ d) culture, excavated in the present Romania, we have started to investigate whether boron and chlorine can be used for the provenance research of high-silica-content rocks. In the Neolithic era, production of polished stone tools started to change the chipped stone industry of the Palaeolithic. A much wider range of volcanic and

108

Zs. Kasztovszky et al.

Fig. 6.7 Reconstruction of distribution of prehistoric obsidian from the Lipari and the Carpathian sources – based on PGAA measurements

metamorphic rocks were used as raw materials. The common term “greenstone” is applied for various metamorphites with different formation. Greenschist, blueschist, Na-piroxenite, serpentinite, nephrite, and hornfels were the most frequently used raw materials for the polished stone artifacts in the Neolithic and the Copper Age period. However, the high-pressure metaophiolites, jadeite, and eclogite were considered as the top-quality raw material. Based on our knowledge about the Carpathian Basin, no geological sources of such raw materials exist in the territory of today’s Hungary. However, there is one exception: one type of greenschist geological source can be found in Felsőcsatár, Western Hungary, and it was possible to identify it with the help of the PGAA measurements [18]. Several national and international projects have been organized to collect reference materials (RMs) from potential sources outside the territory of Hungary, sometimes even outside the Carpathians. During our work, we cooperate with museums in Hungary and in the neighboring countries to investigate their collections. According to a significant result, when comparing the distribution areas of two major raw material types (i.e., of hornfels and contact metabasite), their complementary distribution areas can be seen well. The contact metabasite was found to be dominantly used in the Transdanubian area and in the northeastern parts of Hungary, whereas hornfels was mostly used in the Great Hungarian Plains, in the southeast part of Hungary [43].

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

6.3.2

109

Case Study BNC2: Provenance Research on Archaeological Pottery

Although ceramics is a composite material, it is far from being homogeneous, and therefore its characterization requires the use of several complementary analytical methods. In some fortunate cases, bulk elemental composition data may inform us about the provenance. The manufacturers of the Valencia style pottery from the Lake Valencia Basin (north-central Venezuela mainland) began to colonize the oceanic islands to the north, between A.D. 1200 and the European contact period [44]. Their geographical spread, based on the abundance and size of the archaeological sites of the period, suggests a demographic expansion of the Valencioid people and possibly an effective integration with the local Amerindian populations. Systematic excavations have been made in the Los Roques archipelago, 140 km away from the coast, where the Valencioid sites have been located on six islands [45]. Four sites yielded hundreds of human pottery figurines in primary, ritual contexts [46]. We have performed nondestructive prompt-gamma activation analysis of the figurines to evaluate the hypotheses that might explain the origin of their stylistic differentiation with the help of the compositional data. We expected that the available stylistic and contextual descriptions of the figurines will provide an interpretive frame for the patterns observed in the chemical data. We were able to determine all the major components, such as H2O, Na2O, MgO, Al2O3, SiO2, K2O, CaO, TiO2, MnO, and Fe2O3, with sufficient precision. In addition, we identified some geologically important trace elements, including B, S, Cl, Sc, V, Cr, Ba, Sm, Eu, and Gd with extremely high sensitivities for B and some rare-earth elements. In order to see any clusters among the investigated objects, characteristic elemental ratios were sought, and also the principal component analysis (PCA) was applied to the standardized data [47]. As an outcome of our study, Na2O, K2O, and TiO2 among the major components, as well as Cr and Cl from the trace elements, were found to be distinctive between the samples from the Los Roques and Lake Valencia Basin regions, respectively. The samples from the Los Roques Islands show higher Cl content, which required further investigations. Since these artifacts were buried for 600–800 years and the pottery could have been in contact with seawater during the transportation of the artifacts, the high Cl content could be explained by the adsorption of NaCl from seawater. However, contrary to the behavior of Cl, no such elevation in Na content can be observed at the same time. Also, PGAA provides the average bulk compositions of the samples, which means that the inhomogeneities on the surface are suppressed. The separation of the samples of different origins seems to be reinforced by the principal component analysis, too (see Fig. 6.8). To conclude, based on the PGAA measurements, we can state with high confidence that the ceramic figurines found in the continental region of the Valencia Lake Basin seem to be made of different raw materials from those found in the Los Roques archipelago.

110

Zs. Kasztovszky et al.

3 AM 393

VLB 8843

SOIL 10 VLB 8843 MA 1836 MA 711 VLB 20517

Valencia Lake Basin

2

Los Roques

MA 690a

MA 1572 AM 418

2nd Principal Component

AM 552

1

AM 508

MA 467 VLB 58565 VLB0008

VLB 1415

MA 564 VLB 1590

VLB1 D16

VLB 1601 VLB0075 VLB D33

AM 376

MA 358

0

MA 407 VLB00100

-1 AM 476

MA 371

AM 1067d

AM 3502

-2

VLB 1584

AM 374

VLB 20518 VLB 20519

MA 458

-3

AM 402

VLB 58579 MA 960

-4 VLB 58574

-5 -8

-6

-4

-2

0

2

4

6

8

1st Principal Component

Fig. 6.8 Discrimination between ceramic fragments found at two separate archaeological sites – using the principal component analysis on the measured PGAA data

6.3.3

Case Study BNC3: Provenance Research on Historical Glass

Provenance research of glassware has always been a challenging task for historians. Many times, typological features of the pieces are not sufficient for characterization. The knowledge of chemical constituents may provide important complementary data. Archaeometric study of glass mostly aims to identify certain workshops (also called glasshouses) or to discover historical recipes which are often characteristic to one or more workshops or periods [48]. Many different analytical tools can be used in glass archaeometry. However, most of them require sampling of the historical object. Obviously, in case of valuable, unique pieces, the curators seldom allow taking even a micro-sample from them. The absolutely nondestructive methods are highly preferred here, as is true elsewhere in heritage science. As glass is an amorphous material with homogeneous composition, bulk PGAA can provide meaningful data on it. As a first step, the amount of basic constituents, i.e., of SiO2, CaO, Al2O3, Na2O/K2O, and MgO, might help to limit the region or the period of production. Furthermore, trace elements, associated with the basic raw materials, might give information about the sources of the raw material [49]. Among the trace elements, boron has a specific importance in glass studies. Boron can be found in almost all historical glasses, as it is a natural trace element in certain raw materials used for glass production (such as sand or plant ash). In the ancient times (in Mesopotamia, Egypt, Roman times, or the Medieval times), the boron concentration usually does not exceed trace level.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

111

However, from the seventeenth century, boron was occasionally introduced to glass technology as a separate batch constituent – borax. As a very expensive raw material, it was used almost exclusively in certain types of luxury glassware to create its shiny appearance. Although in most cases the concentration of B2O3 did not exceed a few percent and sometimes it was even below 1 wt%, sometimes it might exceed 10 wt%. During the study of Baroque glass, boron concentration data could provide important technological information. On the other hand, boron is a chemical element that is difficult to measure with most of the other analytical methods, especially in a noninvasive way. PGAA is almost exclusively the method that enables us to measure it. In a particular study, 23 historical glass fragments, mainly from postmedieval and Baroque periods, were examined. Most of the objects were excavated in the area of the Old Town in Elbląg (Poland), a few were found in the Old Town in Poznań (Poland), and a few others originated from private collections. The origin of the pieces is unknown in most cases, as both towns of Elbląg and Poznań were important market centers and many of the pieces undoubtedly were from European import [50]. With PGAA, it is possible to identify most of the major components of glass and to classify the objects. The results obtained were found to be in good agreement with electron probe microanalysis (EPMA) results. Trace elements, such as Sb, Sr, Rb, Y, Zr, and Zn – whose concentrations are important to know in glass archaeometry – are usually below the detection limits of PGAA. Considering the determination of boron concentration, PGAA and EPMA proved to be complementary methods. As a final result, for all of our investigated samples, we found the boron content to be below 0.1 weight %, which means that boron was not deliberately introduced into the glass as a separate raw material but only as a natural contaminant of certain batch constituents [50].

6.3.4

Case Study BNC4: Provenance Research on Archaeological Metal

Prompt-gamma activation analysis, as a bulk method, is a useful tool to determine the alloying components of various metals. As neutrons can penetrate more than one centimeter in any metal, the method enables the investigator to explore the original composition of metal objects, independently from surface alterations. In our heritage science practice, mostly copper and silver alloys but occasionally iron objects were studied. In one particular study, a selected collection of Roman silver coins was brought to the Budapest Neutron Centre for PGAA investigations. The hoard of Roman denarii was found in Romanów (Poland). The treasure was estimated to consist of 700 pieces [51]. Most of the coins are denarii minted between 193 and 197 A.D., i.e., dated to the reign of Didius Julianus, Clodius Albinus, and Septimius Severus. The surviving coins were found in a poor condition. It is supposed that they had been in circulation for a long time before being deposited in Romanów. The largest found collection of

112

Zs. Kasztovszky et al.

coins dated back to the reign of Antonius Pius (including denarii of Faustina the Elder), Marcus Aurelius (140–160 A.D.), and Faustina the Younger (138–160 A.D.). There are also a lot of coins that dated back to the times of Marcus Aurelius (160– 176 A.D.) and Commodus (179–180 A.D.). A common practice in these historical periods was to cut the coins for debasement of their original value, but they still remained a medium of payment. The aim of this study was to determine the alloying components and trace elements of the coins and draw archaeological conclusions regarding the raw material, manufacturing processes, and their debasement. A selection of 115 denarii from the regime of Hadrian (117–138 A.D.) up to Septimius Severus (193– 211 A.D.) has been examined by means of the PGAA method. The Ag/Cu ratios of coins were determined from the PGAA spectra, and it was investigated whether any systematic changes in the composition could be found. We have found from the PGAA measurements that the silver content (or equivalently the Ag/Cu ratio) of the coins was systematically decreasing in time. The mass ratios were found to vary between 1.5 and 4.5. It is important to note that the significant decrease of Ag content was impossible to determine by visual observation. Although the evolution of silver content is slight, more extensive debasement was possible to detect in the period of Antonius Pius (138–161 A.D.); it was milder in the reign of Marcus Aurelius and significant again in the period of fights for power (193–197 A.D.). Two specific denarii issued by Hadrian in 138 A.D. and by Marcus Aurelius in 163 A.D. were identified by PGAA which had an extremely low amount of silver, which indicates that they are probably ancient forgeries from the second century A.D. (see Fig. 6.9). 5.0 Antonius Pius (138-161)

4.5 4.0

Faustina I-II. (141-176) Hadrian (119-138)

Kommodus (177-192) Marcus Aurelius (140-180)

Ag/Cu Mass Ratio

3.5 3.0 2.5 2.0 1.5

FORGERY?

FORGERY?

1.0 0.5 0.0 125

135

145

155

165

175

185

195

Date (Year A.D.)

Fig. 6.9 Decreasing Ag content in ancient Roman coins, measured through the Ag/Cu mass ratios by PGAA

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

6.4

The PGAA Facility at the TUM

6.4.1

The Neutron Source and the Neutron Guide to the PGAA Instrument

113

In the years 2006–2007, the new PGAA instrument of the Technical University of Munich (TUM) was set up at the Forschungs-Neutronenquelle Heinz MaierLeibnitz (FRM II) in Garching bei München, Germany [52]. FRM II is a pooltype reactor with a single cylindrical fuel element in the center and is operated with a maximum thermal power of 20 MW [53, 54]. (The fuel element contains of 113 plates with U3Si2 in aluminum 8.1 kg uranium in total with an enrichment of 92.5% U-235.) It had been designed as a state-of-the-art neutron source with an exceptionally high thermal flux of 81014 cm2 s1, and it replaced the old FRM reactor – the so-called atomic egg. The instrumentation of FRM II has continuously grown since its start in 2004. Recently updated facts and figures concerning FRM II and its instruments, reactor cycle dates, and all needful information for the users of the facility are available online at the websites of the FRM II (www.frm2.tum.de) and the Heinz Maier-Leibnitz Zentrum, MLZ (www.mlz-garching.de). (MLZ is a cooperation between Technical University of the Munich (TUM), the HelmholtzZentrum Geesthacht – Center for Materials and Coastal Research (HZG), and Forschungszentrum Jülich (Jülich Research Centre).) The PGAA instrument is situated at a central position in the neutron guide hall west between the new and the old reactor building (see Fig. 6.10). Except for some control electronics, all parts of the instrumentation are housed in a concrete bunker

Fig. 6.10 Schematic top view of the research reactor complex: On the left side is the old FRM – the “atomic egg” – which is currently under decommissioning (as of June 2019); on the right side are the FRM II, the experimental hall, and the new neutron guide hall east; and in the middle is the neutron guide hall west hosting the PGAA instrument. (Graphic: R. Bucher, J. Fridgen, S. Kressierer, P. Link, C. Sturz, TUM)

114

Zs. Kasztovszky et al.

serving as shielding. The reactor’s cold source [55], which contains 12 liters of liquid deuterium at a temperature of 25 K and a pressure of 1.5 bar, supplies 19 of 32 scientific instruments (2019) including all facilities in the guide hall with cold neutrons. (The other instruments use thermal neutrons, hot neutrons, converted fission neutrons, or even neutron-induced positrons. An ultracold neutron source (UCN) is under construction (2019).) The PGAA instrument is connected to the neutron guide “NL4b.” The curvature of the guide (r ¼ 390 m) ensures that fast neutrons and gamma rays cannot directly reach the PGAA instrumentation (energy cutoff in the cold neutron spectrum). The last 5.8 m of the guide before the bunker is elliptically shaped. In total, the instrument-to-reactor core distance is approximately 50 m. The guide consists of borosilicate glass coated with Ni/Ti multilayers [56].

6.4.2

Irradiation Conditions

Depending on the sample size and matrix, the best choice for sample irradiation is either a collimated beam or a focused beam profile. For this purpose, it is possible to switch between two measuring setups: (1) the collimator (Consisting of 15-mm boron-containing plastic plus 170-mm lead. Beam divergence: 1.5  9% horizontal and 1.4  10% vertical.) and (2) the 1.1-m-long elliptical tapering. (Without vacuum. Beam divergence: 6  30% horizontal and vertical.) The change is pneumatically operated. The sample chamber can be evacuated to avoid the signals from the nitrogen, reduce scattering in air, and keep the sample dry. A medium vacuum of 0.3 mbar is reached within a few minutes and is appropriate for the most purposes. If required, a second vacuum pump is also available to lower the pressure to 103 mbar or below to eliminate any significant influences from neutron capture of nitrogen in air. At the irradiation position, the beam flux is ~2109 cm2 s1 (thermal equivalent) with a size of 20
30 mm2 for the collimated beam and ~ 51010 cm2 s1 for a beam cross section of 11
16 mm2 for the focused beam. If necessary, the flux can be reduced by a set of three attenuators with transmission of 51%, 17%, and 6%. The cold neutron spectrum has an average energy of 1.83 meV (6.7 Å wavelength).

6.4.3

Detectors and Spectrometers

Currently (2019), there are two nitrogen-cooled n-type HPGe detectors for PGAA in use. The detector for standard analysis has a relative efficiency of 60% and is connected in anticoincidence mode to a surrounding BGO detector for Compton suppression. It is usually set to cover the whole energy range of prompt-gamma energies (30 keV to 12,000 keV). The second detector is a low-energy detector (LEGe) and offers a good resolution at low energies and is currently set up to a maximum energy of ~2.5 MeV to include the hydrogen prompt peak. The gamma radiation is collimated with changeable lead collimators, and lithium carbonate windows protect the detectors against scattered neutrons. Both PGAA detectors

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

115

and those for the in-beam neutron activation analysis (described later) use Ortec ® DSpec-50 digital spectrometers for signal processing.

6.4.4

Handling the Samples

The automatic sample changer can store 16 samples. It consists of a motor-driven rotatable magazine and a pneumatic lift. The sample holder frames of the size 80
95 mm2 (inner part 70
60 mm2) are made of aluminum and are stringed with 0.28-mm-thick fluorinated ethylene propylene (FEP) strings without any stoichiometric H atoms. Using FEP is a proven measure to minimize the scattering and to reduce the background [57]. Most of the solid samples can directly be fixed between the strings, while granulates, powders, or liquids can be sealed in polytetrafluoroethylene (PTFE) bags. The maximum sample size is approximately 50
50
20 mm3. Using another sample holder for single samples, it is possible to handle sample sizes close to 100
100
100 mm3. For even larger objects (e.g., amphorae, swords), a measurement in air is possible with a rearrangement of the shielding.

6.4.5

Shielding

The shielding of the instrument surrounds the collimator, the irradiation chamber, the lower part of the sample changer, the detectors, and the beam stop of the instrument. It consists of boron-containing plastic and lead bricks. Close to the detectors, the standard lead alloy containing a few percent of antimony is replaced with a lead alloy containing 1.25% tin and 0.05% calcium. Most of the units of the shielding are moveable to change between the different setups, collimators, etc. The PGAA setup is shown in Fig. 6.11. Further descriptions of the PGAA instrument can be found in [58], in [59], and on the instrument’s web page (www.mlz-garching.de/pgaa) with up-to-date information and the contact addresses for the instrument scientists.

6.4.6

Data Acquisition and Evaluation

The data acquisition and the instrument control are based on the NICOS networkbased control system [60]. The spectrum evaluation is performed with the Hypermet software [61]. In addition to the spectrum, the FWHM and energy calibration and the nonlinear correction are necessary as further inputs. The peak area of the Dopplerbroadened boron peak is calculated separately by using the summing function. The Excel VBA code ProSPeRo written by Révay [20] is used in conjunction with the spectroscopic database in [62] to calculate the masses, concentrations, their uncertainties, and the detection limits for the elements. Figure 6.12 shows the detection limit in a typical mineral sample analyzed with the PGAA instrument at

116

Zs. Kasztovszky et al.

Fig. 6.11 Schematic view of the shielding of PGAA. The sample changer is located above the irradiation chamber. (Design and Graphic: S. Thiel, Institute for Nuclear Physics, University of Cologne)

Fig. 6.12 Detection limit of a typical mineral sample. (Graphic: Zs. Révay and R. Müller, TUM)

the MLZ. Corrections for decay, self-absorption, self-shielding, and interferences are implemented. In the cases, when oxygen cannot be detected at all or only with poor statistics, its content can be calculated from the stoichiometry. In this context, the assumption of the highest oxidation number is valid for many archaeometrical materials like rocks, stones, cement, and glasses, but the numbers differ for organic material and are of course zero for pure metals.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

6.4.7

117

Extensions of the PGAA Instrument

Main parts of the current PGAA instrumentation were operated at the Paul Scherrer Institute (PSI) in Villigen, Switzerland (1997–2002), where the SINQ spallation source supplied it with a neutron flux of 1.4
108 cm2 s1. The planning of the reconstruction in Garching involved groups from Cologne, Munich, Lausanne, and Budapest is described in Kudějová et al. [56]. Since its initial operation as a new instrument in 2007, the PGAA instrument setup has been repeatedly extended with sub-setups, i.e., these methods can be combined with classic bulk PGAA.

PGAI/NT Prompt-gamma activation analysis imaging (PGAI) in combination with neutron tomography (NT) is one of the extensions of the PGAA instrument. PGAI allows for spatially resolved PGAA in the mm range; NT provides information about the inner structure of objects. ▶ Chapter 54 of this book describes the method and the setups, including the one in Garching, in detail. In-Beam NAA A separate low-background counting chamber is placed next to the PGAA bunker. It is used for decay measurements for the so-called in-beam neutron activation analysis (ibNAA) [59]. In this context, “in-beam” means that the activation takes place in the neutron beam – in comparison with the standard “in-core” NAA. The chamber contains a Compton-suppressed 30% HPGe detector cooled by an Ortec ® X-Cooler III mechanical cooler. The setup can reach detection limits comparable to in-core NAA in a small research reactor, due to the high flux of the cold neutron beam and a low background which is 0.80 and 0.64 cps with and without reactor operation, respectively (total cps for 30–6000 keV). Because NAA is complementary to PGAA for certain elements, the combination of the two methods increases the number of detectable trace elements. In archaeometry, a typical example for this is provenance analysis –a larger dataset of trace elements can help to get a more precise statistical separation of different geological deposits. Another advantage in comparison with in-core irradiation is that bigger samples or parts of objects can be analyzed. Larger objects can only be measured in the PGAA sample chamber. NDP Neutron depth profiling (NDP) is a method based on the measurement of the energy loss of charged particles in matter to determine the concentration depth profile of certain nuclides. It strongly depends on the matrix and the initial kinetic energy of the charged particle. Ziegler et al. described the idea in the 1970s [63]. The most common reactions with cold neutrons are (n,p) and (n,α) reactions. Due to their high cross sections, the method is very sensitive to boron and lithium. Taking the abundance and the cross section into account, the sensitivity for element analysis with other stable nuclides is orders of magnitude lower; see, e.g., [64]. Parallel with the growing importance of Li-ion battery research, NDP

118

Zs. Kasztovszky et al.

has experienced a revival and many new technical developments. The NDP setup at the PGAA instrument has already been applied in user projects. The technical details are described in [65] and some applications in [66]. The depth and the vertical resolution are strongly dependent on the matrix, e.g., in graphite the maximum depth is about 50 μm, and the resolution near the surface is 100 nm or better. Given the high degree of specialization on certain elements, NDP applications in the field of archaeometry are rare. Examples are the early studies of Li, B, and N in ancient oriental pottery from Fink and Riederer, published in 1981 [67].

Related Methods and Future Developments Not a part of the PGAA instrument, but at the MLZ the classic NAA (in-core) is also available as well as the instrument FaNGaS (fast neutron-induced gamma ray spectrometry) [68] mainly using the (n,n’) reactions. With FaNGaS, it is possible to analyze larger samples and/or samples with a high absorption cross section for cold and thermal neutrons. The instrument was redesigned in 2019. At the PGAA instrument, a supplement for cyclic in-beam NAA of short-lived activation products (e.g., 20F, ground state of 110Ag) is under construction. Besides the precise analysis of the corresponding elements, this will also help to reduce the activation of intermediate activation products like 110mAg allowing for an earlier radiation protection clearance of metal objects with a higher silver content. Furthermore, a combination of in-beam activation and liquid scintillation counting (LSC), e.g., for the analysis of phosphorous, is under development.

6.5

Scientific Applications

The PGAA instrument is involved in the proposal system of the Heinz MaierLeibnitz Zentrum (MLZ). The instrument is available for scientific and industrial studies. (Industrial users have the opportunity of buying beamtime without the obligation to publish the results.) At FRM II, there are three to four reactor cycles with 60 days each, i.e., in total of 180–240 days of operation. Information about the reactor cycles, the deadlines, and the application process for regular and rapid access proposals are provided on the MLZ user office web page (www.mlzgarching.de/user-office). The applications at the PGAA instrument cover a wide field of scientific disciplines. Archaeology that includes archaeometry and cultural heritage is the topic with the highest demand – it accounted 23% of all beamtime proposals in years 2014 to 2017, for instance (see Fig. 6.13). Other fields were nuclear physics (22%), geosciences (13%), chemistry (11%), condensed matter (7%), medicine (7%), material sciences (4%), soft matter (4%), biology (3%), crystallography (1%), and magnetism (1%). Another 4% of the proposals are related to instrument development.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

119

Fig. 6.13 Pie chart diagram of distribution of the number of proposals at PGAA from 2014 to 2017. Others: Soft matter (4%), biology (3%), crystallography (1%), and magnetism (1%). (Data: MLZ user office / Graphic: C. Stieghorst, TUM)

6.6

Cultural Heritage-Related Applications of PGAA at the FRM II (Case Studies)

A few examples of archaeometry and cultural heritage projects from the scientific users at the PGAA instrument at MLZ will be discussed below. The applications cover a wide field of topics, e.g., provenance analysis, conservation-restoration research, test of purity and authenticity, and reverse-engineering.

6.6.1

Case Study TUM1: Chlorine in Archaeological Iron Objects

The fast corrosion of valuable archaeological iron objects is a major problem for the museums. Even though these objects had been buried in the soil for centuries or even longer, some of them are still relatively well preserved. The high chlorine content together with the humidity and the oxygen in air are important factors which lead to a rapid corrosion of the iron objects. Airtight storage of the objects is obviously complicated and expensive. Alternatively, nondestructive removal procedures can also be considered. There are currently two different approaches for removing chlorine from the iron findings: leaching in alkaline aqueous solutions and heating under a reducing atmosphere. Wagner et al. [69] studied the removal effectiveness of the latter option. They used PGAA to measure the chlorine content of samples before and after the heat treatment in a N2/H2 (90/10) welding gas mixture. The outcome of this procedure is shown in Fig. 6.14 for exemplary test pieces cut from a Celtic iron rod annealed for 24 h at three different temperatures. With the highest temperature of

120

Zs. Kasztovszky et al.

Fig. 6.14 Cl concentrations in six test samples before and after annealing at three different temperatures. (Data and graphic from Wagner et al. [69], modified by: C. Stieghorst, TUM)

750  C, it was possible to reduce the chlorine content to 15% of the initial value. The chlorine concentration could be measured with a detection limit of 10 μg/g (Cl mass/ total iron mass) with a relative uncertainty of about 4%. Additional analyses with Mössbauer and X-ray diffraction (not performed at the MLZ) revealed mineral transformations, e.g., the decomposition of the main chlorine carrier akaganeite at 350  C. The leaching method was also investigated with PGAA in context of another user project (not yet published). Independently from the chosen method, it can be concluded that PGAA is suitable to monitor the effectiveness of the removal process. Though there were photos available, the sample surface can look slightly different after the treatment. The removal effectiveness might be also locally different. However, the inhomogeneities in the initial and the final chlorine distribution in combination with the difficulties in the exact sample positioning make the interpretation of the analytical results complicated.

6.6.2

Case Study TUM2: Archaeological Bronzes

The aim of Maróti et al. [31] was to improve PGAA for the analysis of archaeological bronzes. Their major component is copper with a relatively high neutron capture cross section, and it proved to be rather difficult to analyze minor and trace elements in this matrix [70, 71]. Five certified reference materials (CRMs) were measured: a quaternary alloy, a brass, an arsenic-copper alloy, a lead bronze, and a tin bronze. Different variations of bulk PGAA, in-beam NAA, classic (in-core) NAA, and XRF were conducted at the MLZ in Garching and at the BNC in Budapest, and the results of these methods were compared. As a conclusion, the overestimated tin values in samples containing dominantly copper can be corrected for interference with low-intensity gamma peaks. The authors recommend also high-resolution PGAA using a LEGe detector for the low-energy region to improve the analysis of As, Zn, Sb, Fe, and Ag, as well as PGAA with suppressing the low-energy gamma rays,

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

121

which proved to be advantageous especially for the analysis of Pb, Sn, Ni, Mn, and Fe. A further development is the combination with in-beam NAA to increase the number of the detectable elements.

6.6.3

Case Study TUM3: Provenance Analyses

In 2019, there were two ongoing archaeometry projects with a focus on provenance analysis: the studies of Roman Haltern 70 amphorae from ancient Lusitania and Baetica (Costa et al., publication in preparation) and of ancient Roman limestone building materials. The limestone project was originally set up at the Institute of Nuclear Chemistry, Johannes Gutenberg-University of Mainz, Germany; for previous results, see Stieghorst et al. [72]. In both projects, an elemental “fingerprint,” i.e., a dataset of element concentrations, is used for multivariate statistical data analysis. The required data preparation consisting of missing value treatment, logarithmization, and standardization is also described in the publication mentioned above. For a first structuring of the datasets, e.g., principle component analysis (PCA) or hierarchical cluster analysis including Ward’s minimum variance method is recommended. For classification, support vector machines are one of the most effective methods [73]. For ceramics, it may be necessary to take into account the dilution caused by different temper concentrations; see, e.g., [74, 75]. The concentration dataset of the amphorae is a combination of PGAA and in-beam NAA analyses, both performed at the PGAA instrument at MLZ. Pottery has often high enough minor and trace element concentrations to reach many elements by in-beam NAA. It is necessary to conduct in-core NAA for the limestone to reach a sufficient number of elements for precise results of the multivariate analysis; PGAA data allow for a first course allocation to the limestone deposits. Especially a non-provenance, i.e., mismatch of the fingerprints, can be clearly seen in many cases. The advantages of PGAA are the real nondestructiveness and the faster analysis (no waiting times for long-lived nuclides), so that for an observed mismatch time consumptive NAA measurements can be dispensed. Obviously, there is also no need for destructive sampling, which is an important argument for the curators.

6.6.4

Case Study TUM4: Rattle Rings

Kluge et al. combined bulk PGAA with neutron tomography to analyze the so-called rattle rings [76]. (The name derives from the hypothesis that these rings had been worn on a belt to make noise.) These ring-shaped objects of ~3 cm diameter were found in two different Celtic graves in Bavaria, Germany, the first one in the Huglfing in the year 1889 and the second one in Manching in 1973. The first one consists of loam with four regularly arranged copper pins; the second one shows a corroded iron surface. At first, the archaeologists thought that these objects were made in the same way and the loam ring had just lost its iron content due to corrosion

122

Zs. Kasztovszky et al.

during the centuries. The PGAA and the NT revealed that the second ring was empty, i.e., was manufactured in a completely different way. Further analyses on a higher number of these unusual objects are planned for the next reactor cycles to clarify their purpose in ancient times.

6.6.5

Case Study TUM5: Cosmic Impacts

The younger Dryas period in North America began about 12,900 years ago with a rapid decrease of the temperature and continued with glacial conditions for the next 1300 years. Several species disappeared: most of the megafauna including the mammoths as well as the human Clovis culture. The hypothesis that this was caused by a cosmic impact is widely discussed; see, for instance, [77–80]. Such an impact could be the reason of the thin layer of magnetic grains and spherules of micrometeorites. Many different methods were used to analyze samples from this layer. PGAA measurements show a high hydrogen content in the magnetic grains. A possible explanation is that the impact happened in a water-rich area of the Laurentide Ice Sheet. A high Ti/Fe ratio in the grain could be caused by the impact of a titanium-containing object. The REE pattern in the meteorites is very similar to that of samples collected at the Procellarum Lunar KREEP Terrane on the moon. In search for more Dryas-impact traces across the USA, other impact events were also discovered [81, 82].

6.6.6

Further Projects

In the early phase of the installation of the PGAA instrument, a large international collaboration, the so-called Ancient Charm project, was launched [83]. It is based on the early form of PGAI/NT described in ▶ Chap. 54. A few former projects can be found in [84]. Ongoing user projects are, for instance, the analysis of sinter layers in old Roman aqueducts and bathes to understand the ancient water systems in the Mediterranean region, the bulk analysis, and NT of various Roman findings to study manufacturing processes and how the items were used in ancient times and the investigation of gold findings to clarify their origins.

6.7

The PGAA Facility at NIST

Prompt-gamma activation analysis (PGAA) has been practiced at the National Institute of Standards and Technology (NIST) since 1998, originally employing a beam of thermal neutrons [85] with a flux of 3
108 cm2 s1. A second system was installed on a neutron guide delivering low-energy “cold” neutrons in the NIST Center for Neutron Research (NCNR) [86]. This has since been replaced by a third system at the end position of a converging neutron guide with a thermal equivalent flux of 7
109 cm2 s1 [87], a tenfold increase in flux over the second system. A

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

123

pair of neutron beam collimators further defines the beam after exiting the neutron guide to a spot of about 2 cm FWHM at the sample position. Figure 6.15 shows the neutron beam “footprint” taken inside the PGAA sample chamber using a neutron imaging plate. The simulated cold neutron beam spectrum is shown in Fig. 6.16, with a distribution peaked around 4 Å and a sharp cutoff below 2 Å. Gamma ray spectra are acquired by a HPGe detector system positioned at the side of the beam (Fig. 6.17). This is enclosed in a BGO detector for Compton suppression.

Fig. 6.15 Neutron beam footprint taken inside the PGAA sample chamber along the beam path using a neutron imaging plate (D. Turkoglu et al., unpublished)

Fig. 6.16 Simulated neutron beam spectrum at the exit of the neutron guide for PGAA (J. Cook, unpublished)

124

Zs. Kasztovszky et al.

Fig. 6.17 Plan view of the PGAA instrument with neutron and gamma intensity from Monte Carlo simulation and illustration of a small sample and its spectrum. (From the back cover of Analyst, 2017, 142)

The PGAA capabilities can be extended by several hardware enhancements. An example is using a chopper [88] to produce a pulsed cold neutron beam. This enables the separation of the prompt neutron capture gamma rays from delayed gammas from short-lived activation products by synchronizing data collection with the pulse frequency. The short-lived products are measured during periods when the neutron beam is off. More recently, a linear shutter with a 40-ms transition time has been installed, offering an arbitrary ratio of on/off duty cycles [89]. A collimated or (better) focused neutron beam offers improved spatial resolution when examining large samples by PGAA [90]. The application of a neutron lens to characterize oil paint pigments is described in the case studies (Sect. 6.8). Current research and development includes the Compton imaging of prompt-gamma rays for spatial sensitivity [91].

6.7.1

Standards

The raw data collected by PGAA are the counts of gamma ray photons as a function of energy. However, to be useful, this must be converted into the mass fractions of the elements in the sample. The most straightforward way to do this is through the use of a calibration standard containing a known quantity of the element of interest

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

125

which is irradiated and counted under the same experimental conditions. Therefore, the preparation and use of the standards themselves must be considered. The development of standards is one of the main missions of NIST. There are two major approaches to the development of standards for PGAA. One starts with a known amount of a pure sample of the element of interest. The other approach consists of obtaining a representative sample of a material containing the element, e.g., a basaltic rock, and then determining its mass fraction by a set of independent analytical methods. Ideally in the first approach, the standard would be made up of only the element itself, for example, a foil of a metal such as Ti. However, for elements that are chemically reactive like Ca, it is necessary to prepare the standard in the form of a stable stoichiometric compound with other elements, e.g., C and O, to make calcium carbonate. Preparing multielement standards saves irradiation time. The standard can take one of several physical forms: a powder, a metallic foil, an inert matrix such as a glass, or a solution evaporated on filter paper. In the preparation of Standard Reference Materials ® (SRMs) 2452, 2453, and 2454, standards were prepared by degassing a weighed quantity of titanium in vacuum at high temperature, admitting a measured volume of hydrogen to the system, and allowing it to absorb all the gas as it cooled [92]. The amount of hydrogen in these standards was verified gravimetrically and also by neutron incoherent scattering [93]. Standards prepared for other analytical methods are often useful. For example, if archaeological bronze material is being analyzed, alloys prepared for metallurgical applications may be suitable. CRMs such as NIST SRMs are prepared with careful attention to homogeneity, and their compositions are typically determined by several independent methods. NIST defines several levels of RMs. The most generic is the RM. A CRM is characterized by a metrologically valid procedure for one or more specified properties, accompanied by a certificate that provides the value of the specified property, its associated uncertainty, and a statement of metrological traceability. A NIST SRM is a CRM issued by NIST that also meets additional NIST-specific certification criteria and is issued with a certificate or certificate of analysis that reports the results of its characterizations and provides information regarding the appropriate use(s) of the material (NIST SP 260-136). When PGAA is based on primary standards prepared in the laboratory, the analysis of RMs provides a useful independent verification of the results.

6.7.2

Major Element Ratios

Many cultural heritage objects have dimensions that exceed the diameter of the neutron beam and may also have variable thickness. This makes it difficult to determine the exact irradiated mass. In addition, the effective local neutron flux may vary in intensity due to beam hardening and attenuation. Finally, gamma selfabsorption may also vary from point to point. Thus, it is common in PGAA to

126

Zs. Kasztovszky et al.

compensate for these factors by ratioing the counting rate of the element of interest to that of a major element. This should be distributed relatively uniformly throughout the region being irradiated. It should also have a peak in the gamma ray spectrum as close as possible to that of the peak of interest. Finally, it should have a sufficiently high count rate to avoid propagating a significant increase in uncertainty through the ratio calculation. In some cases, the choice of the major element is obvious. For instance, for bronze objects, it would be Cu, and for ceramics it would be Si. An example of the major element ratio approach is given by the PGAA analysis of the jade object illustrated in Fig. 6.18. This is a modern-day replica made of Wyoming nephrite jade of the blade of an ancient Shang dynasty dagger-axe [94]. Nephrite is a variety of the mineral actinolite which has the composition Ca2(Mg, Fe)5Si8O22(OH)2. It can be seen in the cross sections presented in the figure that the thickness of the blade varies significantly. The blade was irradiated by the 2-cm-diameter cold neutron beam at NGD at three positions indicated by the circles in Fig. 6.18. The count rates for the Ca 1940 keV peak are presented in Table 6.1. It can be seen that they vary significantly, with an uncertainty of 21%. However, when ratioed to the major element Si, the resulting values agree closely: the standard deviation of the mean is 0.5%.

Fig. 6.18 Wyoming nephrite jade replica of a blade from a Shang dynasty jade dagger. The three circles indicate spots analyzed by PGAA. These are identified from left to right as tip, middle and tang

Table 6.1 Ca and Si count rates for nephrite blade Blade position Tip Middle Tang Mean

Ca 1.940 MeV cps 208  0.07 350  0.10 336  0.1 298  64

Si 1.271 MeV cps 88.86  0.05 152.0  0.07 145.4  0.07 129  28

Ca/Si cps/cps 2.341  0.002 2.303  0.002 2.312  0.002 2.319  0.016

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

6.7.3

127

Standard-Free Approach

The major element ratio is in units of cps/cps. This can be converted to a mass fraction using standards as described above. However, these may not be required. If the experimental system is well defined, the relation between counting rate and quantity of element in the sample can be calculated [95]. Employing measurements of detector efficiency (using calibration sources as well as PG lines), together with tabulated partial cross sections, this approach provides an independent check on calibration via prepared standards. Once the equivalency is established, one can perform measurements of unknown samples and obtain elemental mass ratios without requiring standards. Prior prediction of counting rate is especially useful in optimizing the measurement of a valuable sample.

6.7.4

Sample Preparation and Mounting

Small samples, with masses of a few hundred milligrams or less and dimensions smaller than the neutron beam, can be packaged in fluorocarbon film and suspended in the beam. This simple situation is largely free of bias from neutron or gamma ray self-shielding and neutron scattering effects. Although standards can be prepared to have small mass, samples may contain significant quantities of neutron absorbers: for example, the mean thermal flux inside a 1-mm sphere of borosilicate glass is only 87% of that in a similar specimen of quartz. Large samples introduce complications. If the sample is not entirely within the neutron beam, the effective sample mass is ill-defined. If the amount of material exposed to the beam is large, then neutron selfshielding and gamma ray self-absorption may be significant. In addition, if large numbers of neutrons are absorbed, the counting rate may be so high that signal losses due to detector dead time and pileup may bias the results. These issues are minimized by the internal standard method, normalizing the quantity of interest to a known matrix element. The issues involved in extending small-sample analysis to intact artifacts are illustrated by a problem addressed at NIST in 1993 [96, 97]. Compressor blades in jet engines were failing, leading in some cases to the loss of aircraft. Hydrogen embrittlement was suspected, so PGAA was used to quantify the hydrogen concentration. An entire damaged blade (10
30 cm, 650 g) was positioned in the PGAA spectrometer such that a portion of the neutron beam irradiated one of several positions around the periphery of the blade. In this geometry, neither the neutron exposure (the product of neutron flux and sample area) nor the sample mass was well defined. However, the effect of these on the signals of all elements in the sample is the same, so the hydrogen-to-titanium mass ratio, the real quantity of interest, was determined by comparison with small-mass standards of the elements. This internal standard approach is widely applicable [98–100]. Mounting a large sample in the beam may call for ingenuity to assure that the neutrons probe the area of interest and that the gamma detector sees that same area. Several laboratories have used a laser parallel with the neutron beam so that

128

Zs. Kasztovszky et al.

alignment can be optimized without concern for radiation exposure to personnel or activation of the object. Three-dimensional scanning of the object followed by 3D printing of a replica [101] safeguards the original artifact when not actually being irradiated.

6.8

Cultural Heritage-Related Applications of PGAA at the NIST (Case Studies)

6.8.1

Case Study NIST1: Application of a Neutron Lens to PGAA of Oil Painting Pigments

Among the cultural heritage questions that were listed in Sect. 6.3 is the identification of ancient production techniques. In the application to oil paintings, this concerns the determination of the set of pigments, or palette, used by an artist for a given painting. This information can be used in understanding the aesthetic effects achieved by the painter. It can also be used to recognize the presence of an earlier image subsequently covered by another image or areas that have been restored or possible forgeries. Since many historically used pigments were based on inorganic minerals, they can be identified through PGAA by their elemental compositions. However, the most significant pigments may be present as only very small spots that may not generate enough signal under normal levels of neutron beam intensity. Therefore, an investigation was made into the feasibility of using a neutron lens to focus cold neutrons to increase the beam intensity on a very small area [103]. This was done at the PGAA instrument located at the previous neutron guide (NG7) where the flux was an order of magnitude lower [104]. A bender lens made of polycapillary fibers was able to deliver a neutron beam of about 1-mm spot size with an intensity gain of 20 over the normal beam. A sample was constructed by painting on filter paper [105]. Separate paints of cobalt blue and cadmium red dark pigments were composed with linseed oil. Three Cd red lines with widths of about 1.0 mm, 0.1 mm, and 2.5 mm were painted on Whatman filter paper. The entire sheet was then covered with Co blue paint, such that the Cd red lines were invisible. A schematic of this sample is shown in Fig. 6.19, together with the curve created from it [106]. The sample was scanned across the focused beam at 100-μm step size, and the Cd 558-keV gamma peak was integrated as a function of location. The 230-keV Co gamma peak was also monitored, which varied across the scan due to the nonuniformity of the blue paint and neutron attenuation by the Cd. The result suggested that the lens could resolve features greater than 1 mm in size for elements with a large gamma cross section such as Cd. This could potentially be useful to study painting or ink artifacts for art conservation purposes.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

A

unguided beam

guided beam background

B

129

C

70.6 77.5 72.3

67.6 213

77.7

5.0 4.4

3.9

Co blue paint

Cd red paint strips

180 160

net 558 keV eps

140 120 100 80 60 40 20 0 -8.00

-6.00

-4.00

-2.00

0.00

2.00

4.00

6.00

Position (mm)

Fig. 6.19 Left, a neutron focusing lens operated at NG7 for small spot studies [102]. Right: schematic drawing of a sample of paints on filter paper (top), where the blue paint conceals the Cd strips underneath, shown roughly to scale with the x-axis at the bottom. The sample was scanned across the focused beam at a 0.1-mm step size. The Cd peak response as a function of position (bottom curve, from Swider thesis) to resolve the concealed Cd strips spatially

130

Zs. Kasztovszky et al.

Fig. 6.20 Van Dyck painting after restoration (removed on either side 25 cm of later addition) [https://www.nga. gov/collection/art-objectpage.1231.html#history]

In the same work [103], focused beam PGAA was also used to analyze samples of Trondheim stained glass to discern the metallic stains and the bare glass by the varying compositions among B, alkaline (Na, Ca, K, Mg), halogen (Cl, F), and metal (Fe, Cu, Co). The thesis work also examined removed portions of a van Dyck painting (Fig. 6.20), originally painted in ca. 1632. About 25 cm added on either side of the painting was thought to have been added sometime in the eighteenth or nineteenth centuries to fit a new decorative dimension. Small pieces of the removed portion were analyzed in the focused beam PGAA to try to match the paint pigment type such as vermilion, smalt, and ultramarine to the known analysis of the original painting. The most interesting finding is the presence of Cd which was not available commercially as pigments until the mid-nineteenth century, which supports the notion that the addition was made centuries later. The above studies reported results in terms of prompt-gamma peak count rates and therefore did not provide quantitative information in terms of mass or mass fractions. Nevertheless, it was an example of PGAA being used to study cultural heritage qualitatively.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

6.8.2

131

Case Study NIST2: Chemically Bound Water in Smithsonian Building Stones

In addition to the set of heritage science questions identified in Sect. 6.3, PGAA can serve as a diagnostic tool for the conservation of cultural heritage objects and monuments. In particular, PGAA’s high sensitivity to hydrogen suggests that it could be useful for the nondestructive measurement of moisture stone and brick architecture or sculpture. Moisture can cause damage to works of art or architecture made of stone through a number of mechanisms such as direct chemical reaction, formation of expansive hydrous phases, or freeze-thaw cycles [107]. For example, the weathering of the stone of the Sphinx is caused mainly by its sodium chloride content, which can deliquesce and recrystallize in response to atmospheric moisture cycles [108]. This suggests the possibility of a PGAA system based on a neutron generator that could be used to map the spatial distribution of moisture within a structure and its variation with time in response to changes in environmental conditions. Evaluating the feasibility of PGAA in this application is complicated by the fact that it measures total H, but the water in the stone occurs in several states, free, capillary, and chemically bound, that can play different roles in stone deterioration. The free, or bulk, water is found in the coarse pores >100 nm. This will tend to evaporate when the RH (relative humidity) is below 100%. The capillary water occurs in pores with radii between 5 nm and 100 nm where surface tension becomes important because it reduces the equilibrium vapor pressure. This lowers the RH at which evaporation ceases. This saturation RH decreases with decreasing pore size according to the Kelvin equation [109]. The evaporable water fractions, both bulk and capillary water, are the primary agents of stone deterioration. Finally, the chemically bound water (CBW) includes water of hydration, interstitial water, and the OH ion in crystal structures. This state of water does not cause damage, but it creates an H background that defines the minimum level of detection of evaporable water by PGAA. In order to estimate the performance of the PGAA in measuring moisture in stone, it is necessary to have some data on the relative proportions of the three states of water to be found in typical building stones. The maximum amounts of the bulk water in the coarse and capillary pores can be measured by standard gravimetric methods. However, the conventional method for measuring CBW, thermogravimetric analysis, is prone to large scatter in the data because of the very small sample size, 20 mg, and instrumental issues concerning heating rates and maximum temperatures. Therefore, it was decided to use cold neutron PGAA to determine the CBW of a set of samples of typical building stones. The set of stone samples was selected from the stone types used in various buildings of the Smithsonian Institution in Washington, D.C. (Fig. 6.21). These seven stones and one historic brick represented a wide range of mineralogy and geographical origins.

132

Zs. Kasztovszky et al.

Fig. 6.21 Stone and brick samples from Smithsonian Institution buildings

For the PGAA analysis, powdered samples were oven-dried at 105  C for 24 h to drive out the evaporable water. These were then compressed into pellets and sealed with a Teflon wrap. The measurements were made at the original PGAA station at NG7. For the experimental details, see Livingston et al. [110]. The results of the PGAA measurements are presented in Figs. 6.22 and 6.23. The CBW values fall within a relatively narrow range of 0.2% to 0.6%. To put this into perspective, the data for all three states of water are plotted in Fig. 6.24. It can be seen that the background due to CBW is greater than the range of capillary values for all the stones. Thus, PGAA would not be able to measure moisture changes associated with RH cycles. For the more porous stones and the brick, the maximum bulk water content is one to two orders greater than the associated CBW. This implies that PGAA could be applied to measure the transport of water from precipitation, groundwater, or leaking pipes within structures made of these materials.

6.8.3

Case Study NIST3: Provenance of Chinese Nephrite Jade

Ancient Chinese jade objects were made of nephrite, which is a variety of the mineral actinolite: Ca2(Mg, Fe)5Si8O22(OH)2. This was obtained from sources all over Asia including Xinjiang province, China, Siberia, and the Philippines [111]. Thus, there is great interest in a nondestructive elemental analysis method that can be used to determine the provenance of a jade object using the chemical fingerprint approach discussed in Sect. 6.3. Consequently, a preliminary study of the

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

133

Fig. 6.22 Chemically bound water (CBW)

feasibility of using PGAA to identify sources of the Chinese nephrite was performed at NIST. This study did not involve actual ancient jade objects. Instead, reference specimens of nephrite from three source regions in Asia were obtained on loan from the Sackler Gallery of the Smithsonian Institution (Fig. 6.24) [94]. These had previously been analyzed by electron microprobe (EPMA) by the Department of Mineral Sciences of the Smithsonian’s Museum of Natural History [112]. The specimens were on the order of 1 cm in size, and thus the entire volume of each could be irradiated in the 2-cm cold neutron beam. The samples were simply mounted using Teflon threads in the frame. The irradiation times were 2 h. The gamma ray spectra were analyzed by PeakEasy (version 4.84) [113]. The results are presented as ratios to Si mass in order to compensate for sample size, geometry, gamma ray attenuation, etc., as described in Sect. 6.7.1. The EPMA analyses are presented in Table 6.2 as oxides. PGAA was able to detect all these elements as well as Mg and Al which could not be detected by EPMA and also H, which is part of the nephrite crystal structure. The four elements Fe, Mn, Cr, and Ni have typically been used for identifying the nephrite source regions. These are associated with the accessory mineral chromite (FeCr2O4) which gives the jade its greenish hue [114]. The PGAA values for these

134

Zs. Kasztovszky et al.

Fig. 6.23 All states of water

Fig. 6.24 Nephrite jade Smithsonian Institution reference samples, about 1 cm2 in area

elements are plotted in Fig. 6.25. The variations in the elemental ratios between the different regions are much larger than the uncertainties, which indicates that the PGAA would be a suitable method for sourcing the nephrite. The Xinjiang sample has the lowest values, which implies a low content of accessory minerals. This is consistent with its near white color; see Fig. 6.24.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

135

Table 6.2 EPMA analysis of Smithsonian nephrite samples

Oxide SiO2 K2O CaO TiO2 Cr2O3 MnO FeO NiO MgO + Al2O3a Total a

Taiwan USNM 119446 Mass ratio (%) 57.68 0.02 13.36 0.04 0.41 0.12 4.17 0.38 26.44

Xinjiang USNM 6778 Mass ratio (%) 59.9 0.06 13.29 0.02 0.01 0.05 0.2 0.03 26.44

Siberia USNM 6775 Mass ratio (%) 58.05 0.01 13 0.02 0.11 0.1 4.86 0.15 23.7

100

100

100

This sum was estimated as the remaining balance

Fig. 6.25 PGAA results for nephrite sourcing elements. Mean blade ¼ Wyoming, USA source

The correlations between the PGAA and the EPMA values are presented in Fig. 6.26. They all show a positive linear correlation. For Fe, Mn, and Ni, the slopes at 1.0–1.4 indicate reasonably good agreement. The Cr slope is much larger, but the uncertainties for the data points are also much larger, so the slope could be as low as 2. The lack of a one-to-one correlation is partly due to differences in the sampling volumes. EMPA is a point method and thus would be more affected by local variability caused by the random distribution of chromite grains. The PGAA analysis

136

Zs. Kasztovszky et al.

Fig. 6.26 Comparison of PGAA vs EPMA results

interrogated the entire volumes of these samples and thus would give a closer approximation to the actual bulk value. The complete list of elements detected in these samples by PGAA, in alphabetical order, is Al, B, Ca, Cl, Cr, Fe, Gd, H, K, Mg, Mn, Na, Ni, Si, Sm, Sr, Ti, and Zn. In addition to these elements detected by prompt-gamma analysis, Sc and Co were identified by postirradiation delayed gamma counting. The elements Ca, Si, and H form the actinolite crystal structure and thus would not be useful for provenance identification. The four elements Fe, Mn, Ni, and Cr are typically used in EPMA studies of nephrite provenance, as discussed above, and a fifth element, Zn, has also been used to identify nephrite from Taiwan [115]. Thus, the additional elements that PGAA can provide for nephrite provenance include Al, B, Cl, Gd, K, Mg, Na, Sm, Sr, and Ti. Péterdi et al. (2014) used Al, Cl, K, Na, and Ti in a PGAA study to identify the source of the nephrite of an adze blade excavated in Hungary [116]. Of the remaining elements on the PGAA list, B and Sr are of interest because they substitute systematically in the calcium carbonate crystal structure of dolomite, which is one of the reactants in the metamorphic reaction that forms the nephrite. Boron may be especially useful because of its very high partial gamma yield cross section and distinctive peak shape.

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

137

The results of this investigation of the feasibility of PGAA for identifying the sources of Chinese jade objects are promising. The method can detect 15–17 elements potentially suitable for elemental signatures nondestructively. Moreover, the data are bulk averages rather than spot analyses. However, these findings are based on only one sample each from three sources. Additional PGAA data on replicate samples from a greater number of nephrite sources is required to determine which elements have the greatest statistical power for discriminating among sources.

6.9

Concluding Remarks

We can consider prompt-gamma activation analysis as an extension of instrumental activation analysis (NAA) with a specific experimental setup, using an external beam of neutrons to irradiate the samples. The great advantage of the method is that it does not require any sampling from valuable objects. Furthermore, due to the low-intensity external neutron beam, it does not noticeably modify the composition or structure of the investigated objects neither on a macroscopic nor on a microscopic level. PGAA is applicable to quantify some important elements, such as B or the rare earths, which might be important geochemical tracers, as well as H and Cl, which could be good indicators of the corrosion of ancient objects. In return for these useful features, in the case of many trace elements with a few exceptions, the sensitivity of PGAA is not as high as that of NAA. So it is more applicable to quantify the most major components. In addition, there are other – nonnuclear – methods, for instance, ICP-MS and XRF, with which better sensitivities can be achieved for certain elements. These investigations typically cost less, but they are mostly destructive. To summarize, according to our practice, the best results, i.e., the most useful conclusions, can be obtained, when a combination of complementary methods are applied. Acknowledgments Many of the cultural heritage-related projects at the Budapest Neutron Centre have been supported by the National Research, Development and Innovation Fund of Hungary (Nos. K62874 and K100385), the CHARISMA (EC FP7 Grant No: 228330), and IPERION CH (EC H2020 Grant No: 654028) projects of the European Commission. Zsolt Kasztovszky is thankful to Jesse L. Weil for the careful reading and grammatical correcting of the chapter.

References 1. Révay Zs, Belgya T (2004) Principles of PGAA method. In: Molnár GL (ed) Handbook of prompt gamma activation analysis with neutron beams. Kluwer Academic, Dordrecht/ Boston/New York, pp 1–30 2. Lea DE (1934) Combination of proton and neutron. Nature 133:24 3. Amaldi E, D’Agostino O, Fermi E, Pontecorvo B, Rasetti F, Segrè E (1934) Radioattività provocata da bombardamento di neutroni – VII. Ric Sci 2:467–470

138

Zs. Kasztovszky et al.

4. Greenwood RC (1967) Elemental analysis using the neutron capture gamma-ray technique with a Ge(Li) detector. Trans Am Nucl Soc 10:28–29 5. Lombard SM, Isenhour TL (1968) Neutron capture gamma-ray activation analysis using Lithium drifted germanium semiconductor detectors. Anal Chem 40:1990–1994 6. Comar D, Crouzel C, Chasteland M, Riviere R, Kellershohn C (1969) The use of neutron capture gamma radiations for the analysis of biological samples. In: DeVoe JR (ed) Modern trends in activation analysis (NBS Spec. Pub. 312), vol 1. Nat. Bur. of Stand, Washington, DC, pp 114–127 7. Henkelmann R, Born HJ (1973) Analytical use of neutron-capture gamma- rays. J Radioanal Chem 16:473–481 8. Orphan VJ, Rasmussen NC, Harper TL (1970) Line and continuum gamma-ray yields from thermal neutron capture in 75 elements, report DASA 2570 (GA 10278) (gulf general atomic) 9. Hanna AG, Brugger RM, Glascock MD (1981) The prompt gamma neutron activation analysis facility at MURR. Nucl Instrum Methods 188:619–627 10. Lindstrom RM, Zeisler R, Rossbach M (1987) Activation analysis opportunities using cold neutron beams. J Radioanal Nucl Chem 112:321–330 11. Rossbach M (1991) Multielement prompt γ cold neutron activation analysis of organic matter. Anal Chem 63:2156–2162 12. Coceva C (1994) Il Nuovo Cimento 107:85 13. Phillips GW, Marlow KW (1976) Automatic analysis of gamma ray spectra from germanium detectors. Nucl Instrum Methods 137:525–536 14. Révay Zs, Belgya T, Ember PP, Molnár GL (2001) Recent developments in hypermet-PC. J Radioanal Nucl Chem 248:401–405 15. Révay Zs, Belgya T, Molnár GL (2005) Application of hypermet-PC in PGAA. J Radioanal Nucl Chem 265:261–265 16. Molnár GL, Révay Z, Belgya T (2002) Wide energy range efficiency calibration method for Ge detectors. Nucl Instrum Methods A 489:140–159 17. Yonezawa C (1993) Prompt γ-ray analysis of elements using cold and thermal reactor guided neutron beams. Anal Sci 9:185–193 18. Szakmány Gy, Kasztovszky Zs (2004) Prompt Gamma Activation Analysis (PGAA), a new method in the archaeological study of polished stone tools and their raw materials. Eur J Mineral 16(2):285–295 19. Watkinson D, Rimmer M, Kasztovszky Z, Kis Z, Maróti B, Szentmiklósi L (2014) The use of neutron analysis techniques for detecting the concentration and distribution of chloride ions in archaeological iron. Archaeometry 56(5):841–859 20. Révay Zs (2009) Determining elemental composition using prompt γ activation analysis, Anal Chem 81:6851–6859 21. Lindstrom RM, Révay Zs (2004) Beams and facilities. In: Molnár GL (ed) Handbook of prompt gamma activation analysis with neutron beams. Kluwer Academic, Dordrecht/ Boston/New York, pp 49–50 22. Szentmiklósi L, Belgya T, Révay Zs, Kis Z (2010) Upgrade of the prompt gamma activation analysis and the neutron-induced prompt gamma spectroscopy facilities at the Budapest Research Reactor. J Radioanal Nucl Chem 286:501–505 23. Belgya T, Révay Zs (2004) Gamma-ray spectrometry. In: Molnár GL (ed) Handbook of prompt gamma activation analysis with neutron beams. Kluwer Academic, Dordrecht/ Boston/New York, pp 95–97 24. Kasztovszky Zs, Vaday A, Révay Zs, Belgya T, Molnár GL (2000) Prompt gamma activation analysis of bronze archaeological objects. In: Proceedings of the 32nd international symposium on archaeometry 25. Kasztovszky ZS, Szakmány Gy, Révay Zs, Belgya T, Molnár GL (2000) Investigation of neolithic greenschist polished stone tools from Carpathian-Pannonian Basin, Hungary by prompt gamma activation analysis. In: Proceedings of the 32nd international symposium on archaeometry

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

139

26. Libby WF (1946) Atmospheric Helium three and radiocarbon from cosmic radiation. Phys Rev 69:671–2 27. Nelson DE, Korteling RG, Scott WR (1977) Carbon-14: direct detection at natural concentrations. Science 198:507 28. Aitken MJ, Zimmerman DW, Fleming SD (1968) Thermoluminescent dating of ancient pottery. Nature 219:442 29. Martin PD (1977) Dendrochronology and fossil wood – an attempt at dating by heteroconnection. Ann Econ Soc Civilis 32(2):237–239 30. Biró TK, Kasztovszky Z (2018) Stone artefacts and neutrons – case studies from Hungary. J Archaeol Sci: Rep 19:669–673 31. Maróti B, Révay Zs, Szentmiklósi L, Kleszcz K, Párkányi D, Belgya T (2018) Benchmarking PGAA, in-beam NAA, reactor-NAA and handheld XRF spectrometry for the element analysis of archeological bronzes. J Radioanal Nucl Chem 317:1151–1163 32. Sayre EV, Smith RW (1962) Compositional categories of ancient glass. Science 133:1824– 1826 33. Kilikoglou V, Bassiakos Y, Grimanis AP, Souvatzis K (1996) Carpathian obsidian in Macedonia, Greece. J Archaeol Sci 23:343–349 34. Milić M (2014) PXRF characterisation of obsidian from central Anatolia, the Aegean and central Europe. J Archaeol Sci 41:285–296 35. Yi S, Jwa Y (2016) On the provenance of prehistoric obsidian artifacts in South Korea. Quat Int 392:37–43 36. Kasztovszky Zs, Biró KT, Markó A, Dobosi V (2008) Cold neutron prompt gamma activation analysis a non-destructive method for characterization of high silica content chipped stone tools and raw material. Archaeometry 50(1):12–19 37. Kasztovszky Zs, Maróti B, Harsányi I, Párkányi D, Szilágyi V (2018) A comparative study of PGAA and portable XRF used for non-destructive provenancing archaeological obsidian. Quat Int 468:179–189 38. Kasztovszky Zs, Szilágyi V, Biró KT, Težak-Gregl T, Burić M, Šošić R, Szakmány Gy (2009) Horvát és bosnyák régészeti lelőhelyekről származó obszidián eszközök eredetvizsgálata PGAA-val. Archeometriai Műhely 6(3):5–14. (in Hungarian with English abstract) 39. Kasztovszky Zs, Biró KT, Nagy-Korodi I, Sztáncsuj SJ, Hágó A, Szilágyi V, Maróti B, Constantinescu B, Berecki S, Mirea P (2019) Provenance study on prehistoric obsidian objects found in Romania (Eastern Carpathian Basin and its neighbouring regions) using prompt gamma activation analysis. Quat Int 510:76–87 40. Kabaciński J, Sobkowiak-Tabaka I, Kasztovszky Zs, Pietrzak S, Langer JJ, Biró KT, Maróti B (2015) Transcarpathian influences in the early Neolithic Poland. A case study of Kowalewko and Rudna Wielka sites. Acta Archaeol Carp 50:5–32 41. Kasztovszky Zs, Biró KT Fingerprinting carpathian obsidians by PGAA: first results on geological and archaeological specimens. In: Proceedings of 34th international symposium on archaeometry, Zaragoza, 2004. E-book: http://www.dpz.es/ifc/libros/ebook2621.pdf. Insttitución “Fernando el Católico” (C.S.I.C.) Excma. Diputación de Zaragoza, 2006, pp 301–308 42. Biró KT, Kasztovszky Zs Further studies on grey flint samples, 2009. In: Moreau JF, Auger R, Chabot J, Herzog A (eds) Proceedings of the 36th international symposium on archaeometry, 2–6 May 2006, Québec City, Canada. Cahiers d’archéologie du CELAT, n. 25, Série archéométrie n. 7, pp 143–149 43. Váczi B, Szakmány Gy, Starnini E, Kasztovszky Zs, Bendő Zs, Nebiacolombo FA, Giustetto R, Compagnoni R (2019) High-pressure meta-ophiolite boulders and cobbles from northern Italy as possible raw material sources for “greenstone” prehistoric tools: petrography and archaeological assessment. Eur J Mineral 31(5–6):905–917 44. Antczak A (1999) Late prehistoric economy and society of the islands off the coast of Venezuela: a contextual interpretation of the non-ceramic evidence. PhD dissertation, Institute of Archaeology, University of London, UK

140

Zs. Kasztovszky et al.

45. Sajo-Bohus L, Mackowiak de Antczak MM, Greaves ED, Antczak A, Bermudez J, Kasztovszky Zs, Poirier T, Simonits A (2005) Incipient archaeometry in Venezuela: provenance study of pre-Hispanic pottery figurines, J Radioanal Nucl Chem 265(2):247–256 46. Mackowiak de Antczak M (2000) ‘Idols’ in exile: making sense of prehistoric human pottery figurines from Dos Mosquises Island, Los Roques Archipelago, Venezuela. PhD dissertation, Institute of Archaeology, University of London, UK 47. Kasztovszky Zs, Antczak MM, Antczak A, Millan B, Bermúdez J, Sajo-Bohus L (2004) Provenance study of Amerindian pottery figurines with prompt gamma activation analysis. Nukleonika 49(3):107–113 48. Smith RW (1963) In: Brothwell D, Higgs E (eds) Science in archaeology. Thames and Hudson, London, p 520 49. Sayre EV (1966) Refinement in methods of neutron activation analysis of ancient glass objects through the use of lithium drifted germanium diode counters. In: Proceedings of the 7th International Congress on Glass, paper No. 220 Brussels, New York, Gordon and Beach 50. Kasztovszky Zs, Kunicki-Goldfinger J (2011) Applicability of prompt gamma activation analysis to glass archaeometry. In: Turbanti-Memmi I (ed) Proceedings of the 37th international symposium on archaeometry. Springer Berlin/Heidelberg, pp 83–90 51. Kasztovszky Zs, Panczyk E, Fedorowicz W, Révay Zs, Sartowska B (2005) Comparative archaeometrical study of Roman silver coins by prompt gamma activation analysis and SEM-EDX. J Radioanal Nucl Chem 265(2):193–199 52. Kudejova P, Meierhofer G, Zeitelhack K, Jolie J, Schulze R, Türler A, Materna T (2008) The new PGAA and PGAI facility at the research reactor FRM II in Garching near Munich. J Radioanal Nucl Chem 278(3):691–695 53. Gläser W, Petry W (2000) The new neutron source FRM II. Physica B 276–278:30–32 54. Gerstenberg H, Kastenmueller A (2012) Germany’s most modern research reactor, the Forschungs-Neutronenquelle Heinz Maier-Leibnitz (FRM II). atw Internationale Zeitschrift für Kernenergie 58(10):605–607 55. Gaubatz W, Gobrecht K (2000) The FRM-II cold neutron source. Phys B Condens Matter 276–278:104–105 56. Kudějová P, Materna Th, Jolie J, Türler A, Wilk P, Baechler S, Kasztovszky Zs, Révay Zs, Belgya T (2005) On the construction of a new instrument for cold-neutron prompt gamma-ray activation analysis at the FRM-II. J Radioanal Nucl Chem 265(2):221–227 57. Kudejova P, Révay Z, Kleszcz K, Genreith C, Rossbach M (2015) High-flux PGAA for milligram-weight samples. EPJ Web Conf 95:08002 58. Révay Zs (2015) PGAA: prompt gamma and in-beam neutron activation analysis facility. J Large-Scale Res Facil 1:A20 59. Révay Zs, Kudejova P, Kleszcz K, Söllradl S, Genreith C (2015) In-beam activation analysis facility at MLZ, Garching. Nucl Inst Methods A 799:114–123 60. https://forge.frm2.tum.de/nicos/doc/nicos-master/. Accessed 5 June 2019 61. Fazekas B, Molnár G, Belgya T, Dabolczi L, Simonits A (1997) Introducing HYPERMET-PC for automatic analysis of complex gamma-ray spectra. J Radioanal Nucl Chem 215(2):271–277 62. Révay Zs, Firestone RB, Molnár GL, Belgya T (2004). In: Molnár GL (ed) Handbook of prompt gamma activation analysis – analysis with neutron beams. Springer, Boston 63. Ziegler JF, Crowder BL, Cole GW, Baglin JEE, Masters BJ (1972) Boron atom distributions in ion-implanted silicon by the (n,4He) nuclear reaction. Appl Phys Lett 21:16–17 64. Downing RG, Lamaze GP, Langland JK, Hwang ST (1993) Neutron depth profiling: overview and description of NIST facilities. J Res Natl Inst Stand Technol 98(1):109–126 65. Werner L, Trunk M, Gernhäuser R, Gilles R, Märkisch B, Revay Zs (2018) The new neutron depth profiling instrument N4DP at the Heinz Maier-Leibnitz Zentrum. Nucl Instrum Methods A 911:30–36 66. Trunk M, Wetjen M, Werner L, Gernhäuser R, Märkisch B, Revay Zs, Gasteiger HA, Gilles R (2018) Materials science applications of neutron depth profiling at the PGAA facility of Heinz Maier-Leibnitz Zentrum. Mater Charact 146:127–134

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

141

67. Fink D, Riederer J (1981) Studies of Li, B, and N in ancient oriental pottery, and modem ceramic materials by means of (n,p) and (n,α) spectrometry. Nucl Inst Methods 191:408–413 68. Rossbach M, Mauerhofer E (2015) FaNGaS: fast neutron gamma spectroscopy instrument for prompt gamma signature of inelastic scattering reactions. J Large-Scale Res Facil 1:A32 69. Wagner FE, Gebhard R, Häusler W, Wagner U, Albert P, Hess H, Révay Z, Kudejová P, Kleszcz K (2016) Study of archaeological iron objects by PGAA, Mössbauer spectroscopy and X-ray diffraction. Hyperfine Interact 237:30 70. Maróti B, Szentmiklósi L, Belgya T (2016) Comparison of low-energy and coaxial HPGe detectors for prompt gamma activation analysis of metallic samples. J Radioanal Nucl Chem 310:743–749 71. Söllradl S Developments in prompt gamma-ray neutron activation analysis and cold neutron tomography and their application in non-destructive testing. PhD dissertation, Universität Bern 72. Stieghorst C, Kuhnen H-P, Neumann JP, Plonka C (2018) Provenance analysis of Roman limestone from the Moselle valley via neutron activation: research of the Johannes GutenbergUniversität Mainz (Germany), proceedings: Roman ornamental stones in North-Western Europe, Tongeren, Belgium, 20–22 Apr 2016 73. Baxter MJ (2006) A review of supervised and unsupervised pattern recognition in archaeometry. Archaeometry 48(4):671–694 74. Sterba JH, Mommsen H, Steinhauser G, Bichler M (2009) The influence of different tempers on the composition of pottery. J Archaeol Sci 36:1582–1589 75. Beier T, Mommsen H (1994) Modified Mahalanobis filters for grouping pottery by chemical composition. Archaeometry 36(2):287–306 76. Kluge EJ, Stieghorst C, Wagner FE, Gebhard R, Révay Zs, Jolie J (2018) Archaeometry at the PGAA facility of MLZ – prompt gamma-ray neutron activation analysis and neutron tomography. J Archaeol Sci Rep 20:303–306 77. Firestone RB, West A, Kennett JP, Becker L, Bunch TE, Revay Zs, Schultz PH, Belgya T, Kennett DJ, Erlandson JM, Dickenson OJ, Goodyear AC, Harris RS, Howard GA, Kloosterman JB, Lechler P, Mayewski PA, Montgomery J, Poreda R, Darrah T, Que Hee SS, Smith AR, Stich A, Topping W, Wittke JH, Wolbach WS (2007) Evidence for an extraterrestrial impact 12,900 years ago that contributed to the megafaunal extinctions and the Younger Dryas cooling. PNAS 104(41):16016–16021 78. Firestone RB, West A, Bunch TE (2010) Confirmation of the Younger Dryas boundary (YDB) data at Murray Springs, AZ. Proc Natl Acad Sci U S A 107(26):E105 79. Haynes CV Jr, Boerner J, Domanik K, Lauretta D, Ballenger J, Goreva J (2010) The Murray Springs Clovis site, Pleistocene extinction, and the question of extraterrestrial impact. P Natl Acad Sci 107(9):4010–4015 80. Haynes CV Jr, Boerner J, Domanik K, Lauretta DS, Ballenger JAM, Goreva J (2010) Reply to Firestone et al.: no confirmation of impact at the lower Younger Dryas boundary at Murray Springs, AZ. PNAS 107(26):E106 81. Firestone RB, Revay Z (2012) Annual report 2012, Heinz Maier-Leibnitz Zentrum(MLZ) 82. Wittke JH, Weaver JC, Bunch TE, Kennett JP, Kennett DJ, Moore AMT, Hillman GC, Tankersley KB, Goodyear AC, Moore CR, Daniel IR Jr, Ray JH, Lopinot NH, Ferraro D, Israde-Alcantara I, Bischoff JL, DeCarli PS, Hermes RE, Kloosterman JB, Revay Z, Howard GA, Kimbel DR, Kletetschka G, Nabelek L, Lipo CP, Sakai S, West A, Firestone RB (2013) Evidence for deposition of 10 million tonnes of impact spherules across four continents 12,800 y ago. Proc Natl Acad Sci U S A 110:E2088–E2097 83. Schulze R, Szentmiklósi L, Kudejova P, Canella L, Kis Z, Belgya T, Jolie J, Ebert M, Materna T, Biró KT, Hajnal Z (2013) The ANCIENT CHARM project at FRM II: threedimensional elemental mapping by prompt gamma activation imaging and neutron tomography. J Anal At Spectrom 28(9):1508–1512 84. Kudejova P, Revay Zs, Kleszcz K, Söllradl S, Schulze R (2015) High-flux prompt gamma activation analysis based techniques for non-destructive elemental analysis of cultural heritage objects. Restaurierung und Archäologie 8:115–124

142

Zs. Kasztovszky et al.

85. Mackey EA, Anderson DL, Liposky PJ, Lindstrom RM, Chen-Mayer H, Lamaze GP (2004) New thermal neutron prompt gamma-ray activation analysis instrument at the National Institute of Standards and Technology Center for Neutron Research. Nucl Instrum Methods B226(3):426–440 86. Lindstrom RM, Zeisler R, Vincent DH, Greenberg RR, Stone CA, Mackey EA, Anderson DL, Clark DD (1993) Neutron capture prompt gamma-ray activation analysis at the NIST cold neutron research facility. J Radioanal Nucl Chem 167(1):121–126 87. Paul RL, Şahin D, Cook JC, Brocker C, Lindstrom RM, O’Kelly DJ (2015) NGD cold-neutron prompt gamma-ray activation analysis spectrometer at NIST. J Radioanal Nucl Chem 304(1): 189–193. https://doi.org/10.1007/s10967-014-3635-7 88. Zeisler R, Lamaze GP, Chen-Mayer H (2001) Coincidence and anti-coincidence measurements in prompt gamma neutron activation analysis with pulsed cold neutron beams. J Radioanal Nucl Chem 248:35–38 89. Turkoglu DJ, Chen-Mayer HH (2021) Chopped cold neutron beam activation analysis. IEEE Trans Nucl Sci 68(7):1505–1510. https://doi.org/10.1109/TNS.2021.3080682 90. Chen-Mayer HH, Lamaze GP, Mildner DFR, Zeisler R, Gibson WM (2001) Neutron imaging and prompt gamma activation analysis using a monolithic capillary neutron lens. Anal Sci 17: 1629–1632 91. Chen H, Chen-Mayer HH, Turkoglu DJ, Riley BK, Draeger E, Polf JP (2018) Spectroscopic Compton imaging of prompt gamma emission at the MeV energy range. J Radioanal Nucl Chem 318(1):241–246. https://doi.org/10.1007/s10967-018-6070-3 92. Paul RL, Lindstrom RM (2012) Preparation and certification of hydrogen in titanium alloy standard reference materials. Metall Mater Trans A 43A:4888–4895. https://doi.org/10.1007/ s11661-012-1306-2 93. Chen-Mayer HH, Mildner DFR, Lamaze GP, Lindstrom RM (2002) Imaging of neutron incoherent scattering from hydrogen in metals. J Appl Phys 91(6):3669–3674 94. Livingston R, O’Connor A, LaManna J, Chen-Mayer H, Turkoglu D (2018) Investigation of a simulated Chinese jade and bronze dagger-axe by neutron radiography and prompt gamma activation analysis. J Archaeol Sci Rep 21:99–106. https://doi.org/10.1016/j.jasrep.2018.06.011 95. Turkoglu D, Chen-Mayer H, Paul R, Zeisler R (2017) Assessment of PGAA capability for low-level measurements of H in Ti alloys. Analyst 142:3822–3829. https://doi.org/10.1039/ C7AN01308F 96. Paul RL, Lindstrom RM (1994) Determination of hydrogen in titanium alloy jet engine compressor blades by cold neutron capture prompt gamma-ray activation analysis. In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative nondestructive evaluation, vol 13. Plenum, New York, pp 1619–1624 97. Paul RL, Privett HM III, Lindstrom RM, Richards WJ, Greenberg RR (1996) Determination of hydrogen in titanium alloys by cold neutron prompt gamma activation analysis. Metall Mater Trans 27A(11):3682–3687 98. Isenhour TL, Morrison GH (1966) Determination of boron by thermal neutron activation analysis using a modulation technique. Anal Chem 38:167–169 99. Paul RL (1995) The use of element ratios to eliminate analytical bias in cold neutron prompt gamma-ray activation analysis. J Radioanal Nucl Chem 191:245–256. https://doi.org/10.1007/ BF02038220 100. Paul RL (2018) Improving accuracy, precision, detection limits, and sample throughput in prompt gamma-ray activation analysis using cold and thermal neutrons and element ratios. J Radioanal Nucl Chem 318(3):2273–2278. https://doi.org/10.1007/s10967-018-6212-7 101. Szentmiklósi L, Maróti B, Kis Z, Janik J, Horváth LZ (2019) Use of 3D mesh geometries and additive manufacturing in neutron beam experiments. J Radioanal Nucl Chem 320(2):451– 457. https://doi.org/10.1007/s10967-019-06482-0. (in press) 102. Chen-Mayer HH, Mildner DFR, Sharov VA, Xiao QF, Cheng YT, Lindstrom RM, Paul RL (1997) A polycapillary bending and focusing lens for neutrons. Rev Sci Instrum 68(10):3744– 3750

6

Prompt-Gamma Activation Analysis and Its Application to Cultural Heritage

143

103. Swider JR (1998) Focused cold neutron prompt γ-ray activation analysis for the site-specific, Nondestructive Investigation of Cultural Artifacts, Dissertation, University of Maryland, College Park 104. Chen H, Sharov VA, Mildner DFR, Downing RG, Paul RL, Lindstrom RM, Zeissler CJ, Xiao QF (1995) Prompt gamma activation analysis enhanced by a neutron focusing capillary lens. Nucl Instrum Methods B95:107–114 105. Swider JR, Walters WB Focused cold neutrons in prompt-gamma ray activation analysis for the site-specific, nondestructive, investigation of cultural artifacts. Paper presented at the 5th internat. conf. non-destructive testing, microanalytical methods and environmental evaluation for the study and conservation of works of art, Budapest, 24–28 September, 1996 106. Swider JR, Walters WB (2004) Evaluation of sensitivity and resolution enhancement in cold neutron prompt γ-ray activation analysis using neutron lenses. Nucl Instrum Methods B226: 659–666 107. Steiger M, Charola AE, Sterflinger K (2011) Chapter 4: Weathering and deterioration. In: Siegesmund S, Snethlage R (eds) Stone in architecture: properties, durability, 4th edn. Springer, Heidelberg, pp 227–316 108. Gauri KL, Chowdhury AN, Kulshreshtah NP, Punuru AR (1990) Geologic features and durability of limestones at the Sphinx. Environ Geol Water Sci 16(1):57–62 109. Hall C, Hoff WD (2012) Water transport in brick, stone and concrete, 2nd edn. Spon Press, London 110. Livingston RA, Al-Sheikhly M, Grissom C, Aloiz E, Paul R Feasibility study of prompt gamma neutron activation for NDT measurement of moisture in stone and brick. Paper presented at the 40th annual review of progress in quantitative NDE, Baltimore, Maryland, USA, 21–26 July, 2013 111. Hung H-C, Iizuka Y, Bellwood P, Nguyen KD, Bellina B, Silapanth P, Dizon E, Santiago R, Datan I, Manton JH (2007) Ancient jades map 3,000 years of prehistoric exchange in Southeast Asia. Proc Natl Acad Sci U S A 104(50):19745–19750. https://doi.org/10.1073/ pnas.0707304104 112. Douglas JG (1996) The study of Chinese archaic jades using non-destructive X-ray fluorescence spectroscopy. Acta Geol Taiwan 32:43–54 113. Rooney B, Garner S, Felsher P, Karpius P (2015) PeakEasy 4.84: LA-CC-13-040. In: Los Alamos National Laboratory 114. Iizuka Y, Hung HC (2005) Archaeomineralogy of Taiwan nephrite: Sourcing study of nephritic artifacts from the Philippines. J Austronesian Stud 1(1):35–81 115. IIzuka YI, Hung H-C, Bellwood P (2007) Noninvasive mineralogical study of nephritic artifacts from the Philippines and surroundings: the distribution of Taiwan nephrite and implications for island Southeast Asian archaeology. In: Douglas JG, Jett P, Winter J (eds) Scientific research on the sculptural arts of Asia: proceedings of the third Forbes symposium at the Freer gallery of art, Archetype Publications 1 Birdcage Walk London SW1H 9JJ 116. Péterdi B, Szakmány Gy, Bendő Zs, Kasztovszky Zs, Biró KT, Gil G, Harsányi I, Mile V, Szilágyi Sz (2014) Possible provenances of nephrite artefacts found on Hungarian archaeological sites (preliminary results). Archeometriai Muhely XI. 4:207–222

7

Neutron Resonance Analysis Methods for Archaeological and Cultural Heritage Applications Peter Schillebeeckx and Hans Postma

Contents 7.1 Introduction to Neutron Resonance Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Nuclear Physics Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.1 Single-Level Breit-Wigner (SLBW) Formalism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2.2 Doppler Broadening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 Neutron Resonance Spectroscopic Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.1 Pulsed White Neutron Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.2 TOF Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Neutron Resonance Transmission Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4.3 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 Neutron Resonance Capture Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.1 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.2 Data Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.3 Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 Archaeological and Cultural Heritage Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

148 152 152 153 155 156 157 161 161 164 166 167 167 169 177 177 181 182

Hans Postma: deceased. P. Schillebeeckx (*) European Commission, Joint Research Centre, Geel, Belgium e-mail: [email protected] H. Postma Department of Applied Physics, RD&M, Delft University of Technology, Delft, The Netherlands © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_7

145

146

P. Schillebeeckx and H. Postma

Abstract

The use of neutron resonance analysis (NRA) as a nondestructive analysis (NDA) method to determine the overall (bulk) composition of materials is discussed. This can be done by detecting prompt γ-rays, which are emitted after a neutron capture reaction in the object being studied. This technique, known as neutron resonance capture analysis (NRCA), is sensitive to almost all stable nuclides and can be applied to determine the elemental and isotopic composition including trace elements and impurities. It is extensively applied at the time-of-flight (TOF) facilities GELINA and ISIS to study objects and artifacts of archaeological and cultural heritage interest. Another technique, referred to as neutron transmission analysis (NRTA), is based on a measurement of the transmission of neutrons through the object. This is an absolute method that works well for the main elements present in the sample. It is shown that both NRCA and NRTA give consistent and accurate results of bulk compositions. Abbreviations

ANNRI BGO CENDL CERN CSNS ENDF FGM FWHM GELINA ICPS INES JANIS JEFF JENDL J-PARC JRC KURRI LANSCE MLF NAA ND NDA NEA NMA NOBORU NRA

Accurate neutron-nucleus reaction measurement instruments Bismuth germanium oxide Chinese evaluated nuclear data library Conseil européen pour la recherche nucléaire (european organisation for nuclear research) China spallation neutron source Evaluated nuclear data file Free gas model Full width at half maximum Geel linear accelerator Inductively coupled plasma spectrometry (The) Italian neutron experimental station Java-based nuclear data information system Joint evaluated fission and fusion nuclear data library Japanese evaluated nuclear data library Japan proton accelerator research complex Joint research centre Kyoto university research reactor institute Los alamos neutron science center Materials and life science experimental facility Neutron activation analysis Neutron diffraction Nondestructive analysis Nuclear energy agency National museum of antiquities (RMO) Neutron beam-line for observation and research use Neutron resonance analysis

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

NRCA NRTA OECD PGA PMT PSND REFIT ROSFOND RPI SLBW TOF YAP YSO

Neutron resonance capture analysis Neutron resonance transmission analysis Organisation for economic co-operation and development Prompt gamma-ray analysis Photomultiplier Position sensitive neutron detector Resonance fit Russian national library of neutron data Rensselaer polytechnic institute Single-level breit-wigner Time of flight Yttrium aluminum perovskite Yttrium oxyorthosilicate

Symbols

B Bin Bout Cin Cout ΔD εc εγ En Eμ F Fμ φ gJ Γ Γγ Γn I J λ L Ld Lt mn mX nj Nμ ηc

Background Background corresponding to a sample-in TOF spectrum Background corresponding to a sample-out TOF spectrum Sample-in TOF spectrum Sample-out TOF spectrum Doppler width Efficiency to detect a capture event Efficiency to detect a γ-ray Incident neutron energy Resonance energy Self-shielding factor Self-shielding correction factor for resonance μ Neutron fluence rate Spin factor Total resonance width Capture (or radiation) width Neutron width Spin of target nucleus Total angular momentum Orbital angular momentum Wavelength Flight path distance Equivalent distance due to neutron transport in target/moderator Equivalent distance due to neutron transport in the detector Neutron rest mass Rest mass of nucleus X Areal number density Net area of resonance μ in a capture spectrum Ratio of efficiencies to detect a capture event

147

148

P. Schillebeeckx and H. Postma

R R(tm,En) R(Lt,En) Sn σ σ σγ σγ σn σn σtot σ tot td tt tm T Teff θD T0 Ts TLB Texp TM Yc Yexp YM Y0 Ym vn Ψ WX

7.1

Scattering radius Response function for neutron TOF measurements Response function for neutron TOF measurements expressed in equivalent distance Lt Neutron separation energy Cross section Doppler broadened cross section Neutron induced capture cross section Doppler broadened capture cross section Neutron elastic scattering cross section Doppler broadened elastic scattering cross section Total cross section Doppler broadened total cross section Time difference between the moment of detection and the moment the neutron enters the detector or sample Time that a neutron spends in the target/moderator assembly Experimentally observed time difference between stop and start signal Target temperature Effective temperature (FGM) Debye temperature Stop signal (TOF measurements) Start signal (TOF measurements) Transmission based on Lambert-Beer law Experimentally observed transmission Theoretical (model) transmission Capture yield Experimentally observed capture yield Theoretical (model) capture yield Primary capture yield Yield due to scattering followed by neutron capture Neutron velocity Voigt function Weight of nucleus X

Introduction to Neutron Resonance Analysis

The use of neutron resonances for the characterization of materials relies on specific properties of neutron interactions with material [1–3]. The probability that a neutron interacts with a target nucleus strongly depends on its kinetic energy. Even small changes in the neutron energy can result in strong variations of a cross section, which is the quantity used to calculate interaction probabilities of neutrons with material.

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

Cross section / barn

103 10

2

10

1

tot n 

65

100 10-1 10-2 10-3

0

1000

105

Cu Cross section / barn

7

2000

3000

4000

Incident neutron energy / eV

5000

10

149

tot n 

4

115

In

103 102 101 100

0

5

10

15

Incident neutron energy / eV

Fig. 7.1 Cross sections as a function of incident neutron energy for neutron interactions with 65Cu (left) and 115In (right). The total cross section σ tot is compared with the one for elastic scattering σ n and for neutron capture σ γ . The cross sections are Doppler broadened using the free gas model (FGM) with T ¼ 293 K (see Sect. 7.2.2). Note: Cross sections are expressed in units of area (1 barn ¼ 1028 m2). Energies in this chapter are mostly expressed in electron volt (eV)

The strong energy dependence of neutron cross sections is illustrated in Fig. 7.1. This figure shows the total cross section as a function of incident neutron energy for interactions with 65Cu and 115In, together with the one for neutron elastic scattering (n,n) and for neutron-induced capture (n,γ). For relatively low neutron energies, the cross sections show sharp peaks known as resonances. They are related to excited levels of the compound nucleus, which is formed when the incident neutron interacts with a target nucleus [4]. The compound nucleus de-excites mostly by emitting a neutron (neutron elastic scattering), by emitting a cascade of prompt γ-rays (neutron capture reaction), or in case of some heavy nuclides, such as 235U and 239Pu also by nuclear fission (neutron-induced fission). After a neutron capture reaction, the compound nucleus decays to the ground state or in some cases a long-lived isomeric state by emitting mostly several prompt γ-rays in a cascade. For light nuclides or nuclides with a proton or neutron number close to a magic shell, the probability to reach the ground state or isomeric state by a single primary transition can be high. For most nuclides, a large number of primary transitions are possible, leading to a multitude of different cascades and resulting in a complex γ-ray emission spectrum. The total number of transitions that is needed to reach the ground state is defined as the prompt γ-ray multiplicity. The complexity of the possible cascades and the corresponding γ-ray emission spectrum depends on the level scheme of the compound nucleus. A γ-ray spectrum following a neutroninduced capture reaction covers an energy region up to the neutron separation energy of the compound nucleus, which is in the order of 8 MeV. Depending on the energy resolution of the γ-ray detectors, individual γ-transitions can be observed in the low- and high-energy region of a spectrum.

150

P. Schillebeeckx and H. Postma

10-1

1

H

10-4 10-4

16

10-7 101

56Fe

O

Capture cross section / barn

10-2 10-5 102 10

53

Cu

75

As

-1

10-4 103 10

0

10-34 10

109Ag

101 10-2 103 10

197

Au

208

Pb

1

10-2 10-1 10

In

0

10-3 104 10

115

-4

10-7

10-1

100

101

102

103

104

105

Incident neutron energy / eV

Fig. 7.2 Comparison of the cross sections for neutron-induced capture reactions with 1H, 16O, 56 Fe, 53Cu, 75As, 109Ag, 115In, 197Au, and 208Pb as target nuclides. The cross sections are Doppler broadened using the FGM with T ¼ 293 K (see Sect. 7.2.2)

Figure 7.1 illustrates that for a given compound nucleus the interaction cross sections exhibit common resonance structures corresponding to discrete excitation levels of the compound nucleus above the neutron separation energy. Resonance characteristics, in particular their energy and strength, are specific for each element or rather nuclide. This is shown in Fig. 7.2, by comparing the cross section of neutron-induced capture reactions for different target nuclides, i.e., 1H, 160, 56Fe, 53 Cu, 75As, 109Ag, 115In, 197Au, and 208Pb. Therefore, experimentally observed resonance structures or profiles can be used as fingerprints to identify and quantify

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

151

nuclides in an object when they are subjected to a neutron beam [1–3]. This is the basis of neutron resonance capture analysis (NRCA) and neutron resonance transmission analysis (NRTA) to determine the elemental and isotopic composition of materials. NRCA refers to an analysis of resonance structured spectra by recording prompt γ-rays that are emitted after a neutron capture reaction as a function of incident neutron energy. NRTA is based on an analysis of characteristic dips in a transmission spectrum that is obtained by measuring the attenuation of a neutron beam passing through a sample. NRCA and NRTA are nondestructive analysis (NDA) techniques that can be applied to study the bulk elemental and isotopic composition of relatively large archaeological and cultural heritage objects [1, 2]. They are noninvasive and do not require any sample taking or preparation compared to other methods such as neutron activation analysis (NAA) and atomic spectroscopy methods. The induced radioactivity is very low and is mostly well below the limit for free release after a short waiting period. NRCA and NRTA are sensitive to nuclides with strong resonance structures in their neutron interaction cross sections. Fig. 7.2 reveals that for light nuclides and nuclides with a magic proton or neutron number, resonance structures appear only at high energies, e.g., 208Pb. For most of the medium and heavy nuclides, resonance structures are present in a broad energy region starting from low energies. In general, the resonance strength and separation between resonances increases with decreasing energy. These features define the sensitivity of the techniques. The sensitivity for NRCA is discussed in [1, 2]. For nuclides with resonances below 10 eV, the detection limit is in the range of parts per million (ppm). When resonances occur in the region between 10 eV and 500 eV, the lowest detectable weight fraction is between 103 and 105. When resonances can only be observed in the keV region, the detection limit is of the order of 0.01 or even less if no overlapping of high-energy resonances occurs. In general, NRCA has a more favorable detection limit compared to NRTA. The difference is roughly a factor 10. An overview of NRTA and NRCA applications is given in [3]. They can be applied for a variety of applications, including archaeological studies [5–17], characterization of reference samples [18] and complex nuclear materials [19–23], detection of drugs and explosives [24–28], thermometry [29–33], and study of fundamental properties of materials [34–37]. Neutron resonance analysis (NRA) techniques, such as NRTA and NRCA, rely on an analysis of experimentally observed resonance structured profiles [1–3]. Such neutron resonance spectroscopic measurements require a pulsed white neutron source, with an energy distribution covering a broad energy region, combined with the time-of-flight (TOF) technique [38, 39]. The experimental conditions and data reduction and analysis procedures are similar to well-established methods and procedures that are applied for neutron cross section measurements in the resonance region, which are described in [39]. The main difference is that experimental cross section data are obtained from measurements with well-characterized homogeneous samples with a regular shape. In addition, their thickness is mostly optimized considering the resonance characteristics of the nuclide under study. Applying

152

P. Schillebeeckx and H. Postma

NRTA and NRCA as an NDA technique to study archaeological and cultural heritage objects, which have mostly an irregular shape and a considerable thickness, is more complex and requires special procedures.

7.2

Nuclear Physics Background

7.2.1

Single-Level Breit-Wigner (SLBW) Formalism

The origin of neutron resonances is well understood. They were first explained by Bohr, who introduced the compound nucleus model [4]. Resonance structured cross sections can be parameterized in terms of resonance parameters by the R-matrix formalism [40], without requiring any information about the internal structure of the nucleus. Several approximations of this formalism have been proposed [41]. Due to its simplicity and widespread use, the SLBW [42] approximation will be briefly discussed. For an isolated single s-wave resonance, with an orbital angular momentum ℓ ¼ 0, the total cross section σtot as a function of the incident neutron energy En in SLBW approximation is expressed as:
 Γ n En  Eμ R0 λ2 Γn Γ λ2 σ tot ðEn Þ ¼ gJ
þ gJ
þ 4πR2 , ð1Þ 2 2 2 2 4π π En  Eμ þ ðΓ=2Þ En  Eμ þ ðΓ=2Þ where λ is the neutron wavelength, gJ is the spin factor, Eμ is the resonance energy, Γ is the total width of the resonance, Γn is the neutron width of the resonance, and R is the effective scattering radius. The spin factor gJ is the ratio of the number of substates of the compound nucleus with total angular momentum J and the number of substates related to the neutron spin ½ and the nuclear spin I of the target nucleus: gJ ¼

2J þ 1 : 2ð2I þ 1Þ

ð2Þ

The resonance width Γ is the sum of the partial widths. A partial width reflects the probability for a specific interaction to occur. In case only neutron capture and elastic scattering is possible, the resonance width is Γ ¼ Γγ + Γn, where Γγ is the capture width, also referred to as the radiation width. The first term in Eq. 7.1 concerns the resonance total cross section, the third term is due to potential scattering, and the second term is the interference between potential and resonance scattering. The effect of this interference creates for an s-wave resonance a dip in the total and elastic cross section, as illustrated in Fig. 7.1 for the 2530 eV resonance for 65Cu. The capture cross section σγ as a function of neutron energy in SLBW is: σ γ ð En Þ ¼

Γn Γγ λ2 : g 4π J
En  Eμ 2 þ ðΓ=2Þ2

ð3Þ

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

153

Resonance parameters (Eμ, Γn, Γγ, gJ) cannot be predicted by theory. They can only be determined by adjusting them in a fit to experimental data [39, 41]. Recommended resonance parameters can be found in evaluated nuclear data libraries like CENDL (CN), ENDF/B (US), JEFF (EU), JENDL (JP), and ROSFOND (RU). The data in these libraries can be retrieved at the JANIS website of the NEA/OECD [43].

7.2.2

Doppler Broadening

The thermal motion of nuclei in an object will broaden the observed resonance profiles. This is known as Doppler broadening. Within the FGM, a Doppler broadened cross section σ can be approximated by [41, 44]: 1 pffiffiffi σ ð En Þ ffi ΔD π



1 ð

0 

dE e 1

E0 En ΔD

2 rffiffiffiffiffi E0 σ ðE0 Þ, En

ð4Þ

with the Doppler width ΔD defined by: ΔD ¼

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 4En kB T , mX =mn

ð5Þ

where mn and mX are the rest mass of the neutron and target nucleus, respectively, kB is the Boltzmann constant, and T is the sample temperature. The Doppler width relates to the full width at half maximum (FWHM) of the Gaussian distribution by: pffiffiffiffiffiffiffiffi FWHM ¼ 2 ln 2ΔD :

ð6Þ

In case of a solid sample, the FGM can be applied by replacing the temperature T with an effective temperature Teff [45]: T eff

  3 3 θD ’ θD coth , 8 8 T

ð7Þ

where ΔD is the Debye temperature. Debye temperatures for some elements together with the corresponding effective temperatures for a pure elemental sample at T ¼ 293 K are reported in Table 7.1 [46]. The effective temperature as a function of sample temperature for different Debye temperatures is shown in Fig. 7.3. This figure reveals that the effective temperature is only slightly higher than the room temperature when the Debye temperature is significantly lower than the room temperature. With the reported NRA experiments performed at room temperature, it is assumed that the effective temperatures in Table 7.1 are valid for most of the elements present in objects of archaeological or cultural heritage interest.

154

P. Schillebeeckx and H. Postma

Table 7.1 Debye temperatures for some elements together with the corresponding effective temperatures for a sample at T ¼ 293 K. The Debye temperatures are taken from [46]

Element D (K) Cs Pb In Th Au U Cd Ag Ta

38 105 108 163 165 207 209 225 240

Teff (K) 293.2 294.8 294.9 297.2 297.3 299.8 300.0 301.1 302.2 303.1 305.0 311.6

Element

D (K)

Teff (K)

V Mg W Mn Al Co Ni Fe Cr

380 400 400 410 428 445 450 470 630

315.7 318.2 318.2 319.4 321.7 324.0 324.7 327.5 353.9 356.7 841.8

103

θD = 600 K

Teff / K

102

θD = 300 K θD = 100 K

θD

θD = 50 K

1K 50 K 100 K 300 K 600 K

101 θD = 1 K

100 100

101

102

103

Sample temperature, T / K

Fig. 7.3 Effective temperature Teff as a function of sample temperature T for different Debye temperatures θD. The data are calculated using Eq. 7. For low temperatures, the effective temperature approaches 3θD/8

The impact of the Doppler effect on the resonance shape is illustrated in Fig. 7.4 by comparing the capture cross section around the 578 eV resonance of 63Cu for a copper target at T ¼ 0 K with the Doppler broadened cross sections for a sample temperature of T ¼ 300 K and T ¼ 1000 K. This figure demonstrates that the

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

Fig. 7.4 Comparison of the Doppler broadened capture cross sections σ γ ðEn Þ for a 63 Cu target nucleus at T ¼ 0 K, 300 K, 600 K, and 900 K sample temperature. The data are shown in the region of the 578 eV resonance of 63Cu

155

800 T 0K 300 K 600 K 900 K

600 / barn

7

400

200

0

570 580 Incident neutron energy / eV

590

Doppler effect results in a broadening of the resonance profile and a reduction of the peak cross section. However, the resonance area will be maintained. A Doppler broadened capture cross section derived from a combination of the FGM and SLBW approximation has a Gaussian shape around the central energy and a Lorentzian shape in the wings. It can be parameterized by: σ γ ðEn Þ ’ σ 0γ Ψðβ, xÞ,

ð8Þ

with the Voigt function Ψ(β,x) defined as: 1 Ψðβ, xÞ ¼ pffiffiffi β π

ð1 dy 1

ðxyÞ 1  e β2 2 1þy

2

ð9Þ

where β ¼ 2ΔD/Γ, x ¼ 2(E-Eμ)/Γ, and y ¼ 2(E’-Eμ)/Γ. Such a profile can be approximated by the pseudo-Voigt function [2], which is a sum of a Gaussian and a Lorentzian contribution.

7.3

Neutron Resonance Spectroscopic Measurements

The first use of NRA to determine the composition of materials has been reported by Priesmeyer and Harz in 1975 [19]. They obtained the areal density of 131Xe, 133Cs, 152 Sm, 235,238U, and 239Pu present in spent nuclear fuel by NRTA using a fastchopper TOF spectrometer, which was installed at a reactor. From measurements with a chopper [47], resonance profiles can be analyzed with an upper energy limit of about 200 eV. Evidently, this limits the list of elements that can be identified and quantified.

156

P. Schillebeeckx and H. Postma

Energy-selective measurements are also possible at quasi-mono-energetic neutron beams produced by charged particle-induced reactions [48, 49]. Neutron energy-dependent data can be realized by changing the energy of the incident charged particle or by changing the emerging neutron emission angle for a fixed incident charged particle energy. Such measurements have been proposed for the detection of explosives and drugs based on experimentally observed transmission patterns [25]. Due to the limited energy resolution, they are not suitable for a detailed analysis of the elemental and isotopic composition. For applications covering a wide range of elements, a pulsed white neutron source covering an energy region from 1 eV to at least 5 keV combined with the TOF technique is required [50]. This can be realized by a neutron source produced by pulsed charged particle beams based on photonuclear reactions [51] or by the spallation process [52–56], both combined with the TOF technique [57]. Most of the archaeological and cultural heritage studies based on NRA have been carried out at the Geel linear electron accelerator (GELINA) [58] of the Joint Research Centre (JRC) in Geel (BE). They were performed in a collaboration between the JRC Geel and the Delft University of Technology (NL). Studies at the Materials and Life Science Experimental Facility of the Japan Proton Accelerator Research Complex (J-PARC/MLF) [52, 59] concentrated mainly on transmission measurements to test the performance of position-sensitive neutron detectors (PSND) [60–62]. The results of these studies are encouraging to exploit NRA as a tomographic technique providing a spatial distribution of the elemental and isotopic composition. Results reported in [63], which were obtained from a combination of NRCA and prompt gamma-ray analysis (PGA) at J_PARC/MLF [59], are promising. The ISIS facility at the Rutherford Appleton Laboratory (RAL) was used to study detectors for NRA imaging applications [16, 64, 65]. These studies were part of the ANCIENT CHARM project [66]. The suitability of ISIS to determine the compositions of archaeological objects in a semiquantitative way by NRCA was tested in [15–17].

7.3.1

Pulsed White Neutron Sources

At electron-based TOF facilities [51], a pulsed electron beam is stopped in a target to produce bremsstrahlung. Neutrons are generated by subsequent photonuclear reactions. The highest neutron output is obtained by stopping a high-energy electron beam in a heavy metal target, i.e., a target with a high atomic number. At GELINA [58], a pulsed electron beam with an average energy of about 100 MeV is stopped in a mercury cooled, depleted uranium target [67]. A compression magnet reduces the pulse width of the electron beam [68]. The compressed pulse structure can be approximated by a Gaussian distribution with a FWHM less than 1.5 ns. At present, the nominal beam current is about 35 μA with an operating frequency of 400 Hz. At J-PARC/MLF [52] and ISIS [53], high-energy pulsed proton beams, with a maximum energy of 3 GeV and 800 MeV, respectively, are used to generate neutrons in a heavy metal target by spallation reactions. The main components of the spallation

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

Fig. 7.5 Comparison of the energy distribution φ(En) of the neutron fluence rate at a moderated neutron beam of GELINA, the INES beam line of ISIS [74], and the ANNRI [75, 76] and NOBORU [77, 78] beam lines of J-PARC/MLF. The distributions are normalized to φ(En ¼ 1 eV) ¼ 1 eV1 s1 cm2 for a neutron energy En ¼ 1 eV

104

 (En) / (s-1 cm-2 ev-1)

7

157

GELINA INES ANNRI NOBORU

102

100

10-2

10-4 10-3

10-2

10-1 100 101 102 Neutron energy / eV

103

104

targets at J-PARC/MLF and ISIS are mercury and tungsten, respectively. The pulsed proton beams at J-PARC/MLF [52] and ISIS [53] have a double pulse structure. Other TOF facilities where NRA could be applied are, e.g., the proton spallation sources at CNSC (CN) [54], LANSCE (LANL, US) [56], n_TOF (CERN, CH) [55], and the electron-based facilities at KURRI (JP) [69] and RPI(US) [70]. An overview of their characteristics can be found in [2, 39]. Neutrons produced by photonuclear and spallation reactions using high-energy charged particle beams have energies in the MeV region. Their energy distribution is not suited for NRA applications. Therefore, a moderator containing hydrogenous material is used to reduce the energy of the neutrons and to produce a white neutron source with an energy distribution covering the low-energy region down to at least 1 eV [71–73]. At GELINA, two water-filled beryllium containers of 3.7 cm thickness, placed above and below the uranium target, are used as moderators [58]. The neutron energy distribution at a moderated beam of GELINA is shown in Fig. 7.5. It is compared with the one at the INES measurement station of ISIS [74] and the ANNRI [75, 76] and NOBORU [77, 78] stations of J-PARC/MLF. Table 7.2 reveals that in standard operation conditions the absolute neutron intensity at the measurement stations of J-PARC/MLF and ISIS is more than three and two orders of magnitude, respectively, higher compared to the one at GELINA. Therefore, they are better suited for imaging and tomographic applications.

7.3.2

TOF Measurements

Applying the TOF technique, the neutron energy is derived from the time t that a neutron needs to travel a given distance L. For low-energy neutrons, the nonrelativistic expression relating the neutron energy En and velocity vn is valid:

158

P. Schillebeeckx and H. Postma

Table 7.2 Nominal operation characteristics for the JPARC/MLF (ANNRI and NOBORU), ISIS (INES), and GELINA TOF facilities. The corresponding neutron intensity at 1 eV and resolution (ΔLt) for En ¼ 100 eV at ANNRI, NOBORU, and INES are listed together with the values for a moderated beam at GELINA Maximum energy Beam power Frequency Flight path distance Neutron intensity at 1 eV Resolution FWHM, ΔLt at En ¼ 100 eV

(MeV) (kW) (Hz) (m) (s1cm2eV1)

ANNRI 3000 500 25 23 2  107 6 cm

NOBORU 3000 500 25 23 1.4  107 5.5 cm

 2 1 1 L En ¼ mn v2n ¼ mn : 2 2 t

INES 800 160 50 23 7  105 4 cm

GELINA 150 3.5 400 12.5 5  103 2 cm

ð10Þ

The time t is obtained from a measured time difference tm between a stop signal Ts and start signal T0: tm ¼ T s  T 0 :

ð11Þ

The start signal is produced by the pulsed charged particle beam producing the neutrons. The stop signal in an NRTA measurement is processed from a neutron detection system placed in the beam. In case of NRCA, it is provided by the detector recording the γ-rays produced in the capture reaction. The measured time tm is the sum of three contributions [39]: tm ¼ t þ tt þ td ,

ð12Þ

where • t is the time that a neutron leaving the target/moderator assembly with velocity vn (or energy En) needs to travel the distance L between the target and detector or sample • tt is the time that the neutron spends in the target/moderator assembly • td is the time difference between the moment of detection and the moment the neutron enters the detector or sample The times tt and td result from stochastic processes. Therefore, they will broaden the observed resonance profiles. To account for this, broadening response functions of the TOF spectrometer are needed. A response function R(tm, En) is the probability that a neutron with energy En is observed with a time difference tm. The latter is referred to as the observed TOF. The response functions R(tm, En) depend on the distributions of the times (T0, Ts, tt, td) and the distance L. The latter is the distance between the facet of the moderator viewing the beam line and the neutron entrance

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

Fig. 7.6 Comparison of a NRCA spectrum derived from capture measurements at a 12.5 m station of GELINA and measurements at the 23 m INES station of ISIS. The comparison, resulting from measurements using the same sample, illustrates the impact of the double pulse structure at ISIS. The figure is based on measurements that are reported in [64]

159

10000 GELINA (12.5 m) INES / ISIS (23 m)

Counts / (1/s)

7

1000

100

2000

4000

6000

Incident neutron energy / eV

facet of the detector or sample. In case of a regular flat sample, this distance can accurately be determined by metric measurements with an uncertainty of less than 1 mm in case of GELINA. Its contribution to the response function can mostly be neglected. The spreads in T0 and Ts are usually lumped together into one distribution. At GELINA this is a normal distribution with a FWHM less than 2 ns for measurements with scintillators. The double pulse structure at ISIS and J-PARC/MLF can be parameterized by a sum of two Gaussian distributions with equal FWHM as shown in [2]. At ISIS these distributions are separated by 320 ns and have a FWHM of 65 ns. At J-PARC they are separated by 600 ns and have a FWHM of 100 ns. Figure 7.6 illustrates the impact of the double pulse structure by comparing a spectrum obtained at a 12.5 m station of GELINA and a 23 m station of ISIS using the same copper sample. The data of Fig. 7.6 were taken from [64]. The time tt is due to the neutron transport in the target/moderator assembly. Its contribution to the response function strongly depends on the neutron energy as discussed in [39]. To account for this effect, an equivalent distance Lt is introduced, which is defined as [39] Lt ¼ vn tt :

ð13Þ

Applying such a transformation of variables results in probability distributions R(Lt,En) that are less dependent on the neutron energy [3, 39]. Examples of such response functions for a moderated neutron beam at GELINA are given in Fig. 7.7. They are for a flight path that forms an angle of ϑ ¼ 0 with the normal to the exit facet of the moderator. The results in Fig. 7.7, which are obtained from Monte Carlo simulations as described in [79], confirm that the response functions expressed in terms of the equivalent distance are almost independent of neutron energy. For neutron energies above 1 eV, the FWHM is about ΔLt ~ 2 cm, almost independent

160

P. Schillebeeckx and H. Postma 0.5 En 0.1 eV 1 eV 10 eV 100 eV 1000 eV

0.4 R(Lt, En) / (1/cm)

Fig. 7.7 Comparison of the response functions R(Lt, En) for En ¼ 1 eV, 10 eV, 100 eV, and 1000 eV. The distributions are for a moderated neutron beam at a flight path of GELINA that forms an angle of ϑ ¼ 0 with the normal to the exit facet of the moderator. The results are derived from Monte Carlo simulations as described in [79]

0.3

0.2

0.1

0.0

5

0

10

15

20

Equivalent distance, Lt / cm

E n = 100 eV

GELINA INES ANNRI NOBORU

0.004

0.002

0.000

0.02

R(t t, En) / (1/ns)

R(tt, En) / (1/ns)

0.006

E n = 1000 eV

GELINA INES ANNRI NOBORU

0.01

0.00 0

1000 Time, tt / ns

2000

0

500 Time, t t / s

1000

Fig. 7.8 Response functions R(tt,En) for a neutron leaving the target/moderator with an energy of 100 eV (left) and 1000 eV (right). The response functions are for measurements at GELINA [79], at the INES station of ISIS [74], and at the ANNRI [76] and NOBORU [78] stations of J-PARC/MLF

of neutron energy. The broadening of the response functions will increase with increasing angle ϑ. For ϑ ¼ 9 and ϑ ¼ 18 , the FWHM becomes ΔLt ~ 2.5 cm and ΔLt ~ 3.5 cm, respectively. A comparison of response functions for TOF measurements at GELINA, J-PARC/MLF, and ISIS in Fig. 7.8 reveals the superior resolution of a TOF spectrometer using a pulsed white neutron source based on photonuclear reactions compared to the resolution at a spallation source. The differences in resolution can be quantified by the average FWHM of the distributions R(Lt,En). For neutrons leaving the moderator with an energy En ¼ 100 eV, they are ΔLt ~ 2.0 cm, 4 cm, 6 cm, and 5.5 cm for GELINA, INES/ISIS [74], ANNRI/J_PARC/MLF [76], and NOBORU/J_PARC/MLF [78], respectively. It should be noted that the neutron

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

161

energy distribution and response functions at ISIS and J-PARC/MLF strongly depend on the moderator configuration. For measurements at a relatively long flight path, the detection of slow neutrons produced by a previous accelerator pulse will create background events. Evidently the contribution of these neutrons, also referred to as overlap neutrons, depends on the frequency of the accelerator and the flight path length. To reduce this background, a 10B or Cd filter can be placed in the beam. A 1-mm-thick Cd filter will practically remove all neutrons with an energy below 0.55 eV. Such a filter is efficient for measurements at, e.g., a 25 m flight path combined with on operating frequency of 400 Hz. Unfortunately, the transmission through a Cd filter will create structures in the TOF spectrum which might overlap with resonances of the nuclides of interest. This can be avoided by using an overlap filter containing 10B, which has a smooth energy-dependent absorption cross section. Its disadvantage is that it attenuates the neutron beam in a broad energy region. The disadvantage of filters can be removed by using a chopper that is synchronized with the pulsed accelerator beam. Such a chopper is installed at the ANNRI and NOBORU beam lines of J-PARC/ MLF and the INES beam line of ISIS.

7.4

Neutron Resonance Transmission Analysis

7.4.1

Experimental Details

NRTA is based on an analysis of resonance dips that are observed when the probability that a neutron passes through an object without any interaction is measured as a function of its energy. This probability or transmission can be obtained from a measurement with and without a sample in the beam. This is the basis of total cross section measurements, which are the most accurate cross section measurements that can be performed [39]. A schematic design of a transmission measurement station installed at GELINA is shown in Fig. 7.9. The station is described in more detail in [80]. It is equipped with a Li-glass scintillator positioned at about 10 m distance from the neutron producing target. Such detectors are mostly used

Fig. 7.9 Schematic representation of a 10 m transmission system installed at GELINA. More details can be found in [80]

162

P. Schillebeeckx and H. Postma

0.1 eV 1 eV 10 eV 100 eV 1000 eV

100

R(L d, E n) / (1/cm)

Fig. 7.10 Response functions R(Ld,En) for a 6.35mm-thick Li-glass scintillator enriched to 95 wt% in 6Li. The response functions are given for neutrons entering the detector with an energy En ¼ 0.1 eV, 1 eV, 10 eV, 100 eV, and 1000 eV. They are normalized to unity for Ld ¼ 0 cm

10-1

10-2 100 eV 1 eV

1000 eV

10 eV

0.1 eV

10-3

0

1

2

3

4

Equivalent distance, L d / cm

for neutron energies below 1 MeV. The neutron transport in the detector will also contribute to the final TOF response function. This contribution, which depends on the size and neutron cross sections of the main detection material, is best expressed in terms of an equivalent distance Ld [39]. Response functions for a neutron with energy 0.1 eV, 1 eV, 10 eV, 100 eV, and 1000 eV are compared in Fig. 7.10. The data were obtained by Monte Carlo simulations for the 6.35-mm-thick Li-glass scintillator enriched to 95 wt% in 6Li that is used at the 10 m station. A similar detector is installed at a 50 m transmission station of GELINA. This station, which is described in [81], was used in [22] to determine the composition of liquid waste from the reprocessing facility in La Hague containing oxygen, sodium, sulfur, iodine, and lead. The experimental transmission Texp is derived from the ratio of a backgroundcorrected sample-in (Cin) and sample-out spectrum (Cout): T exp ¼

Cin  Bin : Cout  Bout

ð14Þ

The background corrections for the sample-in and sample-out spectra are denoted by Bin and Bout, respectively. All spectra in Eq. 7.14 have the same TOF bin structure. They are corrected for losses due to dead time effects and normalized to the same neutron intensity. The background can be parameterized by an analytical function describing the contribution of a time-independent and a sum of time-dependent components [39]. An example of such a parameterization is given in Fig. 7.11. It shows the background resulting from measurements at a 25 m transmission station of GELINA reported in [82]. The background is approximated by: Bðtm Þ ¼ b0 þ b1 eλ1 tm þ b2 eλ2 tm þ b3 eλ3 ðtm þτ0 Þ :

ð15Þ

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . . 10-3

10-5 10

-6

10

-7

10-8 103

10-5 10-6 10-7 10

104

105

Time-of-flight / ns

106

Cin (20 mm) Bin (20 mm) BOut

10-4 Counts / (1/ns)

Counts / (1/ns)

10

10-3

Cin (0.125 mm) Bin (0.125 mm) Bin (20 mm)

-4

163

-8

10-9 103

b0

b3e-λ3(t+τ0)

104

b1e-λ1t

b2e-λ2t

105

106

Time-of-flight / ns

Fig. 7.11 TOF transmission spectra taken with a 0.125 mm (left) and 20 mm (right) Cu sample in the beam. The spectra result from measurements at a 25 m station of GELINA using a fixed Co and Na background filter. For the 20-mm-thick sample, the sample-in spectrum Cin is shown together with the total background and its components parameterized by Eq. 15 and the background for measurements without sample. The sample-in spectrum for measurements with the 0.125-mm-thick sample is compared with its own background and the background for the 20-mm-thick sample

The parameter b0 is the time-independent contribution. The first exponential is due to the detection of 2.2 MeV γ-rays that are produced by neutron capture in the moderator. Unfortunately, the light produced by a 2.2 MeV γ-ray is close to the one produced in the 6Li(n,α)t reaction. Therefore, this contribution cannot be eliminated by simple energy discrimination. The TOF dependence of this component can be verified by a measurement with a scintillator enriched in 7Li with similar dimensions as the 6Li neutron detector. The second exponential reflects the contribution from neutrons scattered inside the measurement station. The last term accounts for the detection of overlap neutrons. These are slow neutrons from previous accelerator cycles with a TOF that is longer than the operation frequency, which is denoted by τ0. This contribution can be estimated from a measurement with a longer accelerator operation frequency. To determine the total background based on Eq. 7.15, measurements are carried out with black resonance filters in the beam. These are samples containing elements with strong resonances. The thickness of the filters is optimized to have a transmission of about 104. Elements that are often used to produce black resonance filters are, e.g., Ag, Co, Na, W, and S. The parameters in the analytical expression are determined by a least squares adjustment to saturated resonance dips in the spectrum due to the presence of the filters. Figure 7.11 shows samplein TOF spectra taken with a 0.125-mm- and 20-mm-thick Cu sample in the beam together with a fixed Co plus Na black resonance filter. The total background and the contributions of the different components are shown for the measurements with the 20-mm-thick sample. The presence of the sample or black resonance filter in the beam will affect the background [3, 39]. This can be observed in Fig. 7.11 by comparing the background

164

P. Schillebeeckx and H. Postma

for the sample-in measurements with the 0.125 mm and 20 mm sample and the background for the measurements with the 20-mm-thick sample and the one for measurements without sample. To account for this effect, fixed background filters, e.g., Co and Na, are placed in the beam to adjust the amplitudes of the background contributions.

7.4.2

Data Analysis

The experimental transmission in Eq. 7.14 is derived from a ratio of TOF spectra. It does not require any additional measurement to estimate the detection efficiency, the energy distribution of the incident neutron beam, and its absolute intensity. Therefore, it is an absolute measurement. In addition, the theoretical transmission TLB through a homogeneous sample obtained in a good transmission geometry can be directly related to the areal densities and the total cross sections of the nuclides present in the sample, through the Lambert-Beer law: T LB ðEn Þ ¼ e



P j

n j σ tot,j ðEn Þ

,

ð16Þ

whereσ tot,j is the Doppler broadened total cross section and nj the areal number density or number of atoms per unit area of nuclide j present in the sample. A good transmission geometry requires that: • all detected neutrons have passed through the sample • neutrons scattered by the sample and collimators are not detected • the sample is placed perpendicular to the incident neutron beam. These conditions can be realized by a proper collimation of the neutron beam at both the sample and detector position, as shown in Fig. 7.9. When a good transmission geometry is realized and the experimental transmission is obtained after proper correction for beam intensity fluctuations and for dead time and background effects, the areal densities nj can be obtained by a least squares fit minimizing the expression [3]:

T
 χ 2 n j , j ¼ 1, . . . , n ¼ T exp ðtm Þ  T M ðtm Þ V 1 T exp T exp ðtm Þ  T M ðtm Þ ,

ð17Þ

where V 1 T exp is the covariance matrix of the experimental data. The theoretical estimate of the transmission TM(tm) is given by: ð T M ðtm Þ ¼ Rðtm , En ÞT ðEn ÞdE n ,

ð18Þ

with T(En) calculated based on the Lambert-Beer law in Eq. 7.16. The response function R(tm,En) is a convolution of the response due to the neutron transport in the target moderator assembly (Figs. 7.7 and 7.8) and the one due to the transport in the

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

165

detector (Fig. 7.10). The least squares adjustment can be performed by a resonance shape analysis code such as REFIT [83, 84]. This code includes modules to calculate Doppler broadened cross sections starting from resonance parameter and to account for experimental effects such as the response function due to the target/moderator assembly and the neutron detector. An example of such a procedure is shown in Fig. 7.12. The data are obtained from blind test measurements reported in [85]. They were carried out at the 10 m transmission station of GELINA, shown in Fig. 7.9. An artificial sample was produced by a stack of homogeneous metal disks. Each disk was made of a pure element with a stable natural abundance. The areal densities resulting from a fit to the experimental transmission and the reference values are compared in Table 7.3. For

Texp

TM

1.0

Transmission

0.8 0.6 0.4 0.2 0.0

W

Co

Au

Mn

100

Mn

1000

Incident neutron energy / eV

Fig. 7.12 Experimental and calculated transmission as a function of incident neutron energy denoted by Texp and TM, respectively. The data result from measurements at a 10 m transmission station of GELINA with a sample consisting of a stack of metal disks specified in Table 3 [85]. The calculated transmission results from a least squares adjustment with the JRC Geel version of REFIT [84]. The position of pronounced dips due to the presence of Au, Co, Mn, and W is indicated by an arrow

Table 7.3 Comparison of the declared areal number density (nD) and the one derived from NRTA measurements (nNRTA) at a 10 m transmission station of GELINA. The ratio (nNRTA/nD) is given in the last column. The uncertainties on the NRTA data are due to propagating only counting statistics uncertainties and are quoted at 1 standard deviation Element Mn Co W Au

Areal number density (at/b) nD 1.901 (2)  102 4.585 (5)  103 1.337 (1)  103 6.844 (7)  103

nNRTA 1.886 (2)  102 4.550 (66)  103 1.334 (2)  103 6.862 (5)  103

nNRTA/nD 0.992 (2) 0.992 (15) 0.998 (2) 1.003 (1)

166

P. Schillebeeckx and H. Postma

all elements, the agreement is within the uncertainties due to counting statistics. In the adjustment, it is supposed that the nuclear data, i.e., resonance parameters, are known. Evidently, the results strongly depend on the quality of the nuclear data in the data libraries as discussed in [82, 86]. The resonance energies require mostly a slight adjustment.

7.4.3

Perspectives

A study of a flat piece from a bronze vessel, a cauldron from the seventh century BC found in Satricum about 60 km south of Rome, is one of the few examples using NRTA for archaeological and cultural heritage studies based on a quantitative analysis. The composition derived from NRTA is compared with results from NRCA in [2, 87]. The limited use of NRTA is probably due to the irregular shapes of the objects studied so far with NRA. The basic expression of the Lambert-Beer law in Eq. 7.16 is only valid for a homogeneous sample with a constant thickness seen by the incident neutron beam. Most of the resonance shape analysis codes rely on the basic Lambert-Beer law and, as such, cannot be used for studies of irregular and inhomogeneous objects. The accident at the Fukushima Daiichi nuclear power plants triggered an interest in NRTA as an NDA technique to analyze complex inhomogeneous nuclear materials with irregular shapes [3, 23]. Harada et al. [88] reported analytical expressions that can be applied for arbitrary-shaped homogeneous samples with a variable thickness in the direction of the neutron beam. Analytical models to describe the transmission through a stochastic medium, such as a sample consisting of a mixture of particle and rock-like debris of melted fuel, have been proposed and validated at GELINA by Becker et al. [89]. The model includes a parameter to account for the portion of the neutron beam that does not traverse the sample, also referred to as the holes fraction. The implementation of a general model to account for the holes fraction and inhomogeneous samples by areal density probability distributions, as proposed by Kopecky et al. [90], should open an interest in NRTA for a wide range of applications. The effectiveness of the model to account for the holes fraction is demonstrated in [85] and illustrated in Fig. 7.12. One of the disks chosen to create the blind sample was a cobalt disk with a hole. By including in the analysis the impact of the holes fraction, the elemental composition derived from the transmission data was in very good agreement with the declared values, as shown in Table 7.3. The development of imaging applications at cold neutron beams using a PSND triggered the interest to apply NRTA using position-sensitive detectors for the determination of spatial dependent elemental and isotopic distributions. A pixelated PSND, formed by an array of 10  10 6Li scintillators mounted with a pitch of 2.5 mm, was developed by Schooneveld et al. [91]. The performance of this detector to produce 2D elemental-sensitive contrast figures was assessed at GELINA [92] and ISIS [65]. The detector was used at the INES station of ISIS to produce 3D

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

167

elemental-sensitive contrast images of a mediaeval disk fibula from the Hungarian National Museum in Budapest [16]. One of the most promising position-sensitive detectors is the one developed by Tremsin et al. [93, 94]. It has a spatial resolution of about 100 μm and a time resolution of about 500 ns in the thermal energy region and 20 ns in the epithermal region. This detector was used at J-PARC [62], LANSCE [94], and ISIS [95], to produce 2D elemental-sensitive contrast figures. At present the use of PSND is limited to the production of elemental-sensitive contrast figures that do not rely on a quantitative analysis of the data. The main challenge to combine the use of NRA and PSDN into a full quantitative tomographic elemental-sensitive method is to account for the impact of neutron scattering in the sample, as discussed in [65].

7.5

Neutron Resonance Capture Analysis

7.5.1

Experimental Details

NRCA relies on the detection of prompt γ-rays emitted after a neutron capture reaction in the sample. The observable of interest is the capture yield, which is the fraction of the neutron beam undergoing an (n,γ) reaction in the sample as a function of the incident neutron energy. A system suitable for NRCA should have a good time resolution and a low sensitivity to neutrons. These requirements are more important compared to the γ-ray energy resolution of the detectors. A low sensitivity to neutrons means that the probability that neutrons produce a signal due to interactions in the sensitive volume of the detector and its environment (e.g., entrance window of photomultiplier, detector housing) is low. These requirements are very similar to those for detection systems used for capture cross section measurements of non-fissile nuclides. C6D6-based scintillators are considered as one of the best for such studies [39, 96] and have been used for most of the NRCA studies performed at GELINA. To apply NRCA as a routine NDA technique, the analysis routines should not be too complex, and the system should be flexible to accommodate relatively large objects with irregular shapes. The latter condition rules out detection systems covering an almost 4π geometry with a small sample cavity. The performance of different scintillator types, i.e., BGO, BaF2, YAP, and YSO for NRCA applications, considering C6D6 as a reference was assessed in [64]. The time resolution of all these systems is sufficient to study resonance structures by TOF in an energy region below 5 keV. The BGO and BaF2 scintillators suffer from neutron sensitivity in the energy region below 500 eV due to resonances of Ge and Ba. The performance of YAP and YSO crystals in the energy region below 1 keV is comparable to the one of C6D6. Above 1 keV, they suffer from neutron sensitivity due to interactions with yttrium. The TOF facility GELINA is extensively used to study objects of archaeological and cultural heritage interest [5–14]. Three measurement stations at 12.5 m, 30 m, and 60 m distance from the neutron producing target are available. The moderated neutron beams at these stations are collimated to about 7.5 cm diameter at the sample

168

P. Schillebeeckx and H. Postma

Fig. 7.13 Capture measurement setup at a 12.5 m station of GELINA. The setup consists of two C6D6 detectors and an ionization chamber loaded with 10B. The latter is used to determine the energy distribution of the incident neutron beam. The sample is a Cu metal disk used for calibration measurements. Photo by P. Schillebeeckx

position. The stations are equipped with similar detection systems consisting of a set of C6D6 liquid detectors (10 cm in diameter and 7.5 cm length). The scintillators are coupled to a boron-free quartz windowed photomultiplier (PMT). The signals from the PMT are processed by conventional analog electronics, including a timing device to derive the TOF. Frisch-gridded ionization chambers loaded with thin layers of 10B are used to determine the energy distribution of the incident neutron beam based on the 10B(n,α) reaction. The capture system installed at the 12.5 m measurement station is shown in Fig. 7.13. The use of this system for accurate capture cross section data is discussed in, e.g., [97, 98]. An example of a TOF spectrum from measurements of a prehistoric sword (dirk) at the 12.5 m station of GELINA is shown in Fig. 7.14. It results from measurements reported in [14]. Such a TOF spectrum can be viewed during a measurement allowing an online control of the measurement time. This figure, in which the detector response is plotted as a function of TOF, illustrates one of the advantages of NRCA compared to other neutron active interrogation techniques such as PGA. In a capture TOF spectrum, a TOF region can be selected in which the detection of γ-rays due to the presence of impurities or trace elements, such as indium in Fig. 7.14, is dominating and is not disturbed by the presence of other elements. In a conventional PGA measurement, the γ-ray spectrum will be completely dominated by the detection of γ-rays resulting from neutron capture by elements with the largest macroscopic cross section.

7

Neutron Resonance Analysis Methods for Archaeological and Cultural. . .

169

Geistingen Axe

Cu, 230 eV

Counts / (1/ns)

63

350

115

In, Eλ = 1.46 eV

Cexp

300

250

65

Sn, 111.2 eV

700000

850000

105

Ag, 5.2 eV

Sb, 6.2 eV

121

0 104

800000

109

Sb, 21.4 eV Ag, 16.3 eV

123

Sn, 38.8 eV

117

116

500

750000

Time-of-flight / ns

107

65

Counts / (1/ns)

1000

Cu, 1362 eV

Cexp

Cu, 579 eV

x 102

106

Time-of-flight / ns Fig. 7.14 TOF spectrum resulting from measurements with a C6D6 detection system installed at a 12.5 m station of GELINA. The spectrum is taken with a prehistoric bronze axe in the beam. The insert shows the profile around the 1.45 eV resonance of 115In. For some strong resonances, the corresponding nuclides are indicated. The data are based on the measurements reported in [13]

7.5.2

Data Analysis

The energy-dependent capture yield Yc is the sum of primary events Y0,j and multiple interaction events Ym,j for each nuclide j present in the sample [3, 39]: Y c ðEn Þ ¼

X


 Y 0,j ðEn Þ þ Y m,j ðEn Þ :

ð19Þ

j

For a homogeneous slab of material placed perpendicular to a parallel incident neutron beam, the primary yield Y0,j can be expressed as: Y 0,j ðEn Þ ¼ FðEn Þ n j σ γ,j ðEn Þ,

ð20Þ

where σ γ,j ðEÞ is the Doppler broadened capture cross section and nj is the areal density for nuclide j present in the sample. The self-shielding factor F(En): 

P

n j σ tot,j ðEn Þ

1e j Fð En Þ ¼ P , n j σ tot,j ðEn Þ j

ð21Þ

170 0.02

Yield

Fig. 7.15 Theoretical capture yield for a 5-mm-thick Cu disk placed in a parallel neutron beam as a function of incident neutron energy as calculated for a TOF experiment. The results are shown around the 402 eV resonance of 63Cu

P. Schillebeeckx and H. Postma

402 eV

YMC

0.01

414 eV

0.00 390

400

410

420

430

Neutron energy / eV

accounts for the attenuation of the neutron beam in the sample. The reduction in neutron intensity will be strong at the resonance energy and will reduce toward the wings of a resonance. Therefore, it will change the resonance profile. Only for thin samples and/or small cross sections, such that n j σ tot,j 65 [93, 96]. The characteristic X-ray photon emissions in the 1–20 (max 25) keV energy range are detected using high-resolution silicon semiconductor detectors [71]. Protons having energies above the Coulomb barrier excite nuclei in light atoms. Excited nuclei are source of characteristic prompt γ-ray emissions in the 100 keV – a few MeV energy range which are detected using high-resolution germanium detectors [71]. The PIGE spectroscopy is often used together simultaneously with the PIXE as a complementary method. The PIXE is mostly used for analysis of elements with atomic number Z  13, while the PIGE is for analysis of light elements with 3  Z  13. High sensitivity and high precision of the PIXE or PIGE microscopy allow for elemental analysis of the smallest samples like material flakes or grains (2 mm) of rock art paints adhered to limestone were scanned, revealing the pigment’s iron-based composition [6]. Historic objects such as Australia’s oldest European maritime relic, the de Vlamingh plate, and a pewter plate “calling card” left by Dutch explorers on the Australian west coast in 1697 were analyzed at the XFM beamline to help provide details of its corrosion history [81]. The study showed that the 32 cm diameter plate did not require chemical treatments to prevent further corrosion, and revealed that iron nails first were used to mount the plate to its wooden post and when rediscovered a century later was remounted with copper nails. The XFM beamline has been used to analyze the chemical processing of antique photographs [58]. Besides the expected presence of silver and gold, lead was present which did not appear to be part of the photographic process. Fingermarks containing lead were observed on the photograph. Such an observation would be impossible with a point source measurement, thus demonstrating the strength of the scanning XRF method to help understand the source of an element within the context of the photographic image.

15

Cultural Heritage Project at Australian Nuclear Science and Technology. . .

403

To enable the safe and efficient analysis of large objects such as paintings, a dedicated custom-designed scanning station known the milliprobe has been installed at the XFM beamline [33]. It has the ability to scan objects having a mass up to 15 kg over an area 600 mm  1100 mm. The time required to mount a painting is several minutes. The beam size for the milliprobe is defined by a set of X-ray beam slits and the finest resolution is approximately 50 μm, with standard scans performed at 100 μm pixel size. In summary, large area, high definition scanning XRF is a powerful tool for the elemental analysis of a diverse range of cultural heritage objects. The fact that it can both show fine details and maintain the overall context of the sample is extremely useful and significantly reduces guesswork in the interpretation of the data.

15.2.7 Electron Microscopy The electron microscopy facility at ANSTO falls under the nuclear materials development and characterization (NMDC) platform. This platform provides an integrated multidisciplinary service and expertise, including radioactive materials research and development. The electron microscopy facility provides cutting-edge instrumentation to enable ANSTO staff and visiting scientists to perform world-class research and solve major scientific challenges. Our unique ability to handle radioactive materials is what sets us apart from many other laboratories. The NMDC is Australia’s primary source of know-how in the development, characterization, modeling, and testing of radioactive materials and one-stop-shop for radioactive processing facilities unique to Australia. The ANSTO electron microscopy facility is primarily housed in an awardwinning purpose-built building that is designed specifically to enable the highest possible performance from the instruments. The building provides a stable environment with respect to temperature, humidity, mechanical and acoustic vibrations, and electromagnetic fields. The EM building currently houses four electron microscopes, however, one of these is owned by the University of Wollongong (UOW) as part of a research collaboration effort to enable the operation of this cutting-edge instrument almost two years before its final home is completed. External to the EM building, ANSTO operates another two electron microscopes, one of which is housed in a dedicated active characterization laboratory, used for the examination of highly radioactive materials. The other smaller “desktop”-style instrument makes up part of our metallography department and is mostly used for checking sample preparation before sending the samples over to the full size instruments. We have a dedicated metallography department with very knowledgeable staff that are able to use all the equipment needed to cut, mount, and polish or etch samples no matter what the material. We consider this knowledge and experience when it comes to sample preparation to be invaluable for achieving the best possible results, especially with challenging techniques such as electron backscatter diffraction. The metallography

404

F. Salvemini et al.

department also specializes in optical microscopy with a stereoscope, two microscopes (reflected and transmitted light), and a high-resolution inverted microscope. The scanning electron microscope (SEM) capabilities include two focused ion beam instruments; one of these is designed specifically for the characterization of radioactive materials. This dual beam instrument is unique in that it can perform radioactive isotope analysis using a time-of-flight secondary ion mass spectrometer (TOF-SIMS) in conjunction with the xenon plasma ion beam column. In addition to the TOF-SIMS detector our SEMs are also equipped with energy dispersive spectroscopy (EDS), backscattered electron diffraction (EBSD), and a combination of secondary electron and backscattered electron detectors both chamber mounted and in lens for higher resolution and signal filtering. The SEM allows for nondestructive analysis of samples, which is very advantageous for forensic and heritage materials or anything that cannot be disturbed. The only real prerequisite for materials going inside the SEM is that they are free of moisture to prevent issues under vacuum. Nonconductive samples can also pose a problem in traditional electron microscopes due to sample charging; however, our SEMs are capable of imaging insulating samples through the use of a localized nitrogen gas injection system to remove charging. This gas injection system coupled with the instruments highly optimized electron optics that can operate at extremely low beam energies allow us to examine even the most delicate samples at high resolution. To be able to perform X-ray analysis in a nondestructive manor on insulating materials will require a conductive carbon coating due to the requirement for the higher beam energies to generate the X-rays. Another area of expertise within the platform is the fabrication of micromechanical testing samples using the ion beam and then performing in situ experiments at high resolution including tension, compression, cantilever bending, and fatigue mode. The transmission electron microscope (TEM) capabilities include two instruments, however, as mentioned above one of these is owned by the UOW. The ANSTO-owned TEM is used by researchers to study a range of materials. Of particular interest are materials that could be used in generation four nuclear reactors. The TEM is a 200 keV system capable of diffraction contrast imaging; electron diffraction patterns; atomic resolution imaging; scanning transmission electron microscopy (STEM) imaging; energy dispersive spectroscopy (EDS) for chemical analysis of elements heavier than C; and energy filtered TEM (EFTEM) for chemical analysis of lighter elements. The UOW-owned instrument is a Krios G3i Cryo-TEM primarily used for the study of biological materials. This TEM enables life science researchers to unravel life at the molecular level. It will enable new ways of understanding and curing diseases such as motor neurone disease, Alzheimer’s and cancer, as well as tackling health challenges like antimicrobial resistance, and could assist developing new targeted cancer treatments. The Krios has a highly stable 300 kV emission system, with sample autoloader for cryogenic sample manipulation. It can perform automated applications, such as single particle analysis and cryotomography, and has 0.7 Å resolution.

15

Cultural Heritage Project at Australian Nuclear Science and Technology. . .

15.3

405

Examples from the Project at ANSTO

This section showcases a selection of successful case studies where the application of a multi-technique approach demonstrated to be paramount in shading light on the most diverse scientific questions.

15.3.1 Forensic Analysis of Ancient Swords Filomena Salvemini, Vladimir Luzin, Maxim Avdeev, Sue Gatenby, Min-Jung Kim, Francesco Grazzi.

Scientific Background Today regarded as work of art, the Samurai sword dominated the battlefield of Japan over thousands of years, combining balance, artistic beauty, and sharpness in its cutting edge. These distinctive aesthetical and mechanical features were obtained through a unique manufacturing process that involved specific thermomechanical treatment of a layered structure of steel components with different carbon content. Over time, the laminated composite structure differentiated and evolved among traditions and schools. The crafting methods were never documented, and the necessary information was orally transmitted from the master to his most skilled pupils. Despite the large amount of studies published on the subject, different manufacturing techniques are still not fully understood [53]. In the past, investigations were conducted by using invasive point-based methods requiring sample cutting or by probing limited areas of the surface; nowadays, the most advanced scientific analytical tools have enabled researchers to examine this ancient weapon in a complete non-invasive way. Neutron imaging and diffractions methods allow studying the composition and microstructural properties imprinted by the manufacturing process. This approach can provide a more holistic understanding of ancient manufacturing processes by investigating the bulk material while preserving the integrity of the object [68]. In this study, neutron methods were exploited to develop a multi-technique analytical protocol to scientifically attribute Japanese swords of unknown origin by characterizing their bulk structure as a result of the manufacture. The reader is referred to chap. 55 Cultural Heritage Applications at the J-PARC where a more in-depth description of the traditional manufacturing process and illuminating examples of the application of neutron methods to the study of the Samurai sword are presented in detail by Prof. Yoshiaki Kiyanagi. Materials and Methods The object of our investigation is a set of four katana from the East Asian Collection of the Museum of Applied Arts and Sciences (MAAS) in Sydney (AU) (Fig. 15.11) (Table 15.1). Transliteration of the signature engraved on the hilt defines geographic location, time period, and author of three blades. Manufacture details were not

406

F. Salvemini et al.

H5378 Tsuguhiro Iwami 1661 - 1673

100 mm

Fig. 15.11 Photographic image showing one of the investigated swords. (Adapted from Ref. [123])

Table 15.1 The list of the investigated samples. Period, province, and author are reported for the attributed samples. The unknown katana is here referred to as mumei ID H1360 H5378 H4839 H6856

Type Katana Katana Katana Katana

Period 1346–1370 ACE 1661–1673 ACE 1800 ACE unknown

Province Bitchu Iwami Taizen unknown

Author Sadatsugu Tsuguhiro Yokohama Sukenaga mumei (unknown)

known for one of the samples, so-called mumei (H6856). Despite stylistic analyses are widely adopted to asset the production period, it lacks in rigor, repeatability, and impartiality. A more robust scientific analytical approach based on neutron tomography, diffraction, and residual stress analysis was explored in this forensic study. In particular, neutron tomography was used to characterize the bulk structural and morphological features (i.e., amount, distribution, and shape of defects, porosity, and non metallic inclusions). Neutron diffraction full-pattern analysis was exploited to non-invasively quantify metal and non metal phases composing the samples. A quantitative phase analysis can contribute to the understanding of the smelting and smithing procedures. Neutron diffraction stress analysis was carried out in the predetermined locations of the transverse cross-section of the sample to determine the residual stress as an imprint of the manufacturing process. In order to distinguish unambiguously among possible laminated structures, measurements were carried out with resolution (a size of the projection of the gauge volume into the transverse cross-section) of 0.5  0.5 mm2. At each mesh point, strain measurements were acquired in three orthogonal directions (longitudinal, transverse, and vertical) as in a traditional stress scanning experiment. In the attempt to define the manufacturing technique based on scientific data rather than stylistic criteria, our approach consisted in comparing the distinctive features identified for the set of blades of certain attribution – used as benchmarks – with the features detected in the blade of unknown origin.

Results and Discussions A synthetic explanation of the obtained results will be provided in this section, while a more detailed discussion can be found in ref. [123, 124]. Structural components such as metal, slag inclusions, and porosities were inspected via tomographic analysis (Fig. 15.12). Since each component differs in neutron attenuation, it can be represented as a distinctive intensity value in the

15

Cultural Heritage Project at Australian Nuclear Science and Technology. . .

407

Fig. 15.12 Cross sections along the xy plane (a) and the yz plane (b) of the tomographic reconstruction of sample H5378. At the bottom-right corner, the bar indicates the correspondence between gray tone and linear attenuation coefficient. (Image adapted from Ref. [124])

Fig. 15.13 On the left, preliminary ECHIDNA neutron diffraction data collected for one of the blades showing the major ferrite α-Fe phase; on the right, zoomed in region showing minor phases of cementite Fe3C (blue) and magnetite Fe3O4 (red). (Adapted from Ref. [122])

tomographic reconstruction. A threshold-based segmentation method [117] can be used to partition and quantify each component in order to draw some observation about the extent of refinement of the raw material use in the making of the blade. For example, low-quality items will show high concentrations of defects and inclusions; these will be almost absent in high-quality sword that usually underwent more extensive treatment before forging. Complementary neutron diffraction analysis allowed us to gain information about the crystalline structure of the samples. The smelting process can be inferred from the detection and quantification of wustite, fayalite, and troilite, while the carburization of steel can be quantified from the relative concentration of ferrite and cementite [48]. The concentrations of goethite, magnetite, and hematite can be exploited to asset the oxidation level during the smithing process, as well as during the lifetime of the artifact. An example of a typical diffraction pattern is shown in Fig. 15.13. The most decisive data were provided by residual stress measurements. Two-dimensional maps of the two stress components, d0 and peak width, were obtained for all samples the same way [123], but demonstrated here just for one sample (Fig. 15.14) (In contrast to the traditional stress experiment, where d0 must be provided and three stress components are calculated, in the current experiment, the zero-through-thickness-stress condition was applied enabling calculation of two

408

F. Salvemini et al.

max: 90

Residual stress: L and T, MPa max: 169

ID

250

0.014

d0 maps, Å 1.1744 1.1742 1.174

0.013

150 100

1.1738

0.012

50

0.011

1.1736

0.01

1.1734

0.009

1.1732

-200

0.008

1.173

-250

0.007

1.1728

0.006

1.1726

0.005

1.1724

0.004

1.1722

0 -50 -100 -150

-300

min: -88

min: -447

T

0.015

200

H6856 N

L

Peak width, Å

300

-350 -400 -450 -500

400

Longitudinal stress, MPa

Longitudinal stress, MPa

Fig. 15.14 A generic mesh in the transverse cross-section of the sword for neutron diffraction mapping and a real example of the longitudinal stress d0 and peak width maps for all blades. (Adapted from Ref. [123])

200 0 -200

H5378 H4839

-400 -600 0.0

0.2

0.4

0.6

0.8

1.0

distance from the cutting edge, rel. units

400 200 0 -200

H1360 H6856

-400 -600 0.0

0.2

0.4

0.6

0.8

1.0

distance from the cutting edge, rel. units

Fig. 15.15 Similarities in the central line profiles of the longitudinal stress component allowed identifying two groups of blades. Since the sample differ in size, all curves are plotted in the scale normalized to 1 to make results geometrically comparable. The stresses were measured with an accuracy of 80%

0.10

1

2

3

4

5

6

Sample (# N)

2.15 b)

#5

2.05 #3

P b/

206

Pb

2.10

208

#6

#4

2.00 #2 1.95

#1

0.83

0.84

0.85

0.86 207

0.87

0.88

0.89

0.90

206

Pb/ Pb

another for the Greek coins coming from Byzantium, during the Roman Empire, from the Macedonian Kingdom, and from the Greek Syracuse in Sicily (Italy).

16.4

Discussion and Conclusions

The present research was focused on archaeo-metallurgical characterization of six bronze coins, three found in Alexandria of Egypt dated third to sixth centuries after Christ and three found in different places but coming from the Greek period ranging between fourth before Christ and first after Christ, coming from Greece up to Sicily,

462

L. Torrisi et al.

in Syracuse. After a bibliographic review and a numismatic research, the examined coins were identified as coins used during the Byzantine Empire (#1, #2), Roman Empire (#3, #4), Macedonian Kingdom (#5), and Greek expansion in Sicily, Syracuse (#6). The identification was possible through the good images impressed by the coinage on the two faces of each coin. The samples study was carried out with a multianalytical approach of physical analyses applied to archaeometry, such as the X-ray fluorescence induced by electron beam (EDX), optical microscopy and scanning electron microscopy (SEM), Rutherford backscattering spectrometry (RBS), optical spectroscopy analysis (LIBS), and laser ablation coupled to mass quadrupole spectroscopy (LAMQS). The results of the compositional measurements have shown that the bronze composition is different for the surface patina and for the bulk. A patina surface layer of 100 nm and more in thickness consists of many oxides, overall Cu2O and CuO; chloride compounds, such as FeCl; and sulfur compounds, such as FeS, and other impurities containing C, Mg, Al, Si, Cl, K, and Ca are present. The coin matrix contains high concentration of copper. The coins #4, #5, and #6 have a Cu atomic content higher than 80%. All the bronze alloys contain tin and lead, and, generally, the concentration of Sn is lower with respect to that of Pb. The study of corrosion products with a use of such techniques as EDX, RBS, and LAMQS reveals the presence of oxides, mainly copper oxides that create the red-colored corrosion patina. The presence of chloride compounds indicates the occurrence of a phenomenon which appears in the result of after-burial contact of bronze or other copper alloys with chlorides deposited in a soil or saltwater that contaminate the copper coins. The fingerprint given by the LAMQS analysis of the Pb stable isotopes with respect to the data books of the minerals present in literature indicates possible origin of the minerals from which Cu and Pb were extracted. Such information can be useful to understand the commercial and economic processes during the times of Byzantine and Roman Empires, the Macedonian Kingdom, and Greek colonization. It is possible to hypothesize that the minerals used to coinage the coins of Alexandria of Egypt come from Timna mines and the commerce of this town was extended up to 500 km far from it [25]. Similarly, the Greek expansion and colonization was also developed to transfer minerals of copper and lead from the far Thera mines, in Santorini Isle, up to all the Greece and up to Syracuse, in Sicily colonies to mint Greek money [26]. Thus, presented data can be useful for numismatic, historical, and archaeological interest and should be further developed in order to understand better the metallographic features which can help in determining the mechanical and thermal processing at which the coins were submitted during their manufacturing techniques.

References 1. Torrisi L, Caridi F, Giuffrida L, Torrisi A, Mondio G, Serafino T, Caltabiano M, Castrizio ED, Paniz E, Salici A (2010) LAMQS analysis applied to ancient Egyptian bronze coins. Nucl Instrum Methods B268:1657–1664

16

Mass Quadrupole Spectrometry Coupled to Laser Ablation for Cultural. . .

463

2. Torrisi L, Caridi F, Borrielli A, Giuffrida L, Torrisi A, Mondio G, Mezzasalma A, Serafino T, Caltabiano M, Castrizio ED, Paniz E, Romeo M, Salici A (2010) LAMQS and XRF analyses of ancient Egyptian bronze coins. Radiat Eff Defects Solids Inc Plasma Sci Plasma Technol 165(6):626–636 3. Ding F, Qian Y, Deng Z, Zhang J, Zhou Y, Yang L, Wang F, Wang J, Zhou Z, Shen J (2018) Size-selected silver nanoparticles for MALDI-TOF mass spectrometry of amyloid-beta peptides. nanoscale 10(46):22044–22054 4. Plotnikov A, Vogt C, Hoffmann V, Täschner C, Wetzig K (2001) Application of laser ablation inductively coupled plasma quadrupole mass spectrometry (LA-ICP-QMS) for depth profile analysis. J Anal At Spectrom 11:1290–1295 5. Torrisi L, Italiano A, Torrisi A (2016) Ancient bronze coins from Mediterranean basin: LAMQS potentiality for lead isotopes comparative analysis with former mineral. Appl Surf Sci 387:529–538 6. De Laeter JR, Heumann KG, Rosman KJR (1991) Isotopic compositions of the elements 1989. J Phys Chem Ref Data 20(6):1327–1341 7. Litron Laser, actual website 2019. https://litron.co.uk/product-range/high-energy-compactlasers/ 8. PrismaPlus Pfeiffer-Vacuum, actual website 2019. https://www.pfeiffer-vacuum.com/en/ products/measurement-analysis/analysis-equipment/residual-gas-analysis/residual-gas-analy sis-in-ultra-high-vacuum/prismaplus-qmg-220-m3-1-300-amu/ 9. Torrisi L, Sciuto A, Calcagno L, Musumeci P, Mazzillo M, Ceccio G, Cannavò A (2015) Laserplasma X-ray detection by using fast 4H-SiC interdigit and ion collector detectors. J Instrum 10: 1748. https://doi.org/10.1088/1748-0221/10/07/P07009 10. Torrisi L, Mondio G, Mezzasalma AM, Margarone D, Caridi F, Serafino T, Torrisi A (2009) Laser and electron beams physical analyses applied to the comparison between two silver tetradrachm greek coins. Eur Phys J D54:225–232 11. SREM code, The Stopping and Range of Electrons in Matter, Actual website 2019. http://www. srim.org/SREM.htm 12. Ziegler J (2019) SRIM code, the stopping and range of ions in matter, actual website: http:// www.srim.org/ 13. Messina University (2019) Actual website: https://www.unime.it/it/dipartimenti/dicam 14. Craddock PT (1978) The composition of the copper alloys used by the Greek, Etruscan and Roman civilizations. Part 3. The origins and early use of brass. J Archaeol Sci 5:1–16 15. Catanzaro EJ, Murphy TJ, Shields WR, Garner EL (1968) Absolute isotopic abundance ratios of common, equal-atom, and radiogenic lead isotopic standards. J Res Nat Bureau Standards-A Phys Chem 72A(3):261–267 16. Brettscaife.net geological Database (2019) Actual web-site: http://www.brettscaife.net/lead/ data/index.html 17. OXALID – Oxford Archaeological Lead Isotope Database (2019) Actual website: http://oxalid. arch.ox.ac.uk/Greece/Greece.html 18. Aiello D, Buccolieri A, Buccolieri G, Castellano A, Di Giulio M, Sandra Leo L, Lorusso A, Nassisi G, Nassisi V, Torrisi L (2007) Selective laser cleaning of chlorine on ancient coins. In: Proceedings SPIE 6346, XVI International symposium. Gas Flow and Chemical Laser & High Power Laser Conference Gmunden 4–8 sett 06 19. BAM, Certified Reference Materials, alloys and bronze, actual website. 2019. https://rrr.bam. de/RRR/Content/EN/Downloads/RM/crm-catalogue.pdf?__blob¼publicationFile 20. Mezzasalma AM, Mondio G, Serafino T, DeFulvio G, Romeo M, Salici A (2009) Ancient coins and their modern fakes: an attempt of physico-chemical unmasking. Mediter Archaeol Archaeom 9(2):15–28 21. NIST (2019) Atomic weights and isotopic compositions with relative atomic masses, actual website: https://www.nist.gov/pml/atomic-weights-and-isotopic-compositions-relative-atomicmasses 22. OrlicBacheler M, Biscan M, Kregar Z, Badovinac IJ, Milosevic JD (2016) Analysis of antique bronze coins by laser induced breakdown spectroscopy and multivariate analysis. Spectrochim Acta Part B Atom Spectrosc 123:163–170

464

L. Torrisi et al.

23. Torrisi L, Venuti V, Crupi V, Silipigni L, Cutroneo M, Paladini G, Torrisi A, Havránek V, Macková A, La Russa MF, Birarda G, Vaccari L, Macchia A, Khalilli F, Ricca M, Majolino D (2019) RBS, PIXE, ion-microbeam and SR-FTIR analyses of pottery fragments from Azerbaijan. Heritage 2:1852–1873. https://doi.org/10.3390/heritage2030113 24. Ingo GM, De Caro T, Riccucci C, Angelini E, Grassini S, Balbi S, Bernardini P, Salvi D, Bousselmi L, Ilingiroglu AC, Gener M, Gouda VK, Al Jarrah O, Khosroff S, Mahdjoub Z, Al Saad Z, El-Saddik W, Vassiliou P (2006) Large scale investigation of chemical composition, structure and corrosion mechanism of bronze archeological artefacts from Mediterranean basin. Appl Phys A Mater Sci Process 83:513–520 25. Mann A, Begemann F, Hardheitkemper E, Pernicka E, Schmitt-Strecker S (1992) Early copper produced at Feinan, Wadi Araba, Jordan: The Composition of Ores and Copper Archeomaterials 6:1–33 26. Gale NH, Stos-Gale ZA (1982) Bronze age copper sources in the Mediterranean: a new approach. Science 216(4541):11–19

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to Handheld Instrumentation

17

Giorgio S. Senesi and Olga De Pascale

Contents 17.1 17.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The LIBS Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.1 Principles, Mechanisms, and Processes of Plasma Formation and Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.2 LIBS Instrumentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.2.3 Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3 Applications of Portable and Handheld LIBS Instrumentations to Cultural Heritage Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3.1 Stone Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3.2 Pigments and Mural Paintings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3.3 Metal Objects, Coins, Pottery, and Jewelry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.3.4 Submerged Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17.4 Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

467 468 468 471 473 475 475 482 489 495 500 501

Abstract

Over the last years, laser-induced breakdown spectroscopy (LIBS) has gained a very important role as a tool for in-situ cultural heritage investigations owing to its high sensibility to light elements such as H, Li, B, C, N, and O, ease of use, no need of sample pretreatment, robustness and versatility, noninvasiveness, microdestructivity, and availability of compact transportable setups. More recently, mobile LIBS instrumentations have been developed and improved which allow the performance of trustable, contactless, fast, sensitive, multielemental analysis with a minimum impact on the art objects. In this chapter, a review is provided of the results obtained in-situ, outdoor and indoor in laboratories or museums,

G. S. Senesi (*) · O. De Pascale CNR, Istituto per la Scienza e Tecnologia dei Plasmi (ISTP) – sede di Bari, Bari, Italy e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_17

465

466

G. S. Senesi and O. De Pascale

specifically by the use of mobile, i.e., “portable,” “transportable,” and handheld, LIBS instruments. In particular, LIBS applications to the analysis of monument stones, pigments, mural paintings, metal objects, coins, pottery, and jewelry are reviewed. Finally, new trends and future perspectives of LIBS as an efficient analytical tool to be extended to the analysis of even submerged materials are highlighted. Abbreviations

3D AAS CCD CF CW DP DRS ESA FTIR IB ICCD ICP-OES LDA LIBS LIPS LTE MP MSL NASA Nd:YAG NIR OCT OM PIXE SEM-EDX SIMS SOM SP UV VTS XRD XRF μLIBS μXRF

Three-dimensional Atomic absorption spectroscopy Charge-coupled device Calibration-free Continuous wave Double pulse Diffuse reflectance spectroscopy European Space Agency Fourier transform infrared Inverse Bremsstrahlung Intensified charge-coupled device Inductively coupled plasma-optical emission spectrometry Linear discriminant analysis Laser-induced breakdown spectroscopy Laser-induced plasma spectroscopy Local thermodynamic equilibrium Multi-pulse Mars Science Laboratory National Aeronautics and Space Administration Neodymium-doped yttrium aluminum garnet Near-infrared Optical coherence tomography Optical microscopy Particle-induced X-ray emission Scanning electron microscopy-energy-dispersive X-ray Secondary ion mass spectroscopy Self-organized maps Single pulse Ultraviolet Virtual thin sections X-ray diffraction X-ray fluorescence Micro-laser-induced breakdown spectroscopy Micro-X-ray fluorescence

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

17.1

467

Introduction

Cultural heritage objects consist of an extremely wide variety of materials, structures, and dimensions that make difficult to find an analytical technique optimal in all cases. Most artifacts are unique and precious; thus noninvasive techniques are required for analytical purposes. Further, several objects, such as wall paintings and monument building stones, cannot be transferred to the analytical laboratory because they are located outdoor and/or are portions of large and complex structures. Even in the case the object can be moved to the laboratory for analysis, this procedure is generally in conflict with classical conservation ethics based on avoiding any risk of artifact damage. This implies the need to dispose of mobile, portable instruments for in-situ analysis and, in some cases, such as monument facades, to perform measurements remotely so avoiding any transport constrictions. Over the last years, mobile instrumentations of growing relevance have been developed and improved to allow the performance of trustable, in-situ, contactless, fast, sensitive, multielemental, noninvasive, non- or micro-destructive analysis with a minimum impact on the art object. Thus, sampling, transportation, and any possible alteration of the physical and chemical integrity of the artifact can be avoided, and the risk of accidental damage and related insurance costs can be markedly reduced [1]. This result has been achieved by miniaturizing instrumental components, developing more compact design and improving detector technology, instrument-computer interfacing, focusing optics, and radiation sources [1]. For example, mobile instrumentation has been used for the direct object analysis in archaeological sites and outcrops, in exhibition rooms of museums, and for wall paintings and rock arts. The analyses were finalized not only to conservation and restoration treatments but also focused on the materials and techniques used to build the artwork, in order to obtain chronological and geological information [1]. The mobile analytical instruments most frequently used are based on X-ray fluorescence (XRF) and Raman spectroscopy [1]. Although XRF is a noninvasive technique, its main disadvantage consists in the difficulty to measure in air elements with atomic number below 13, such as C, Na, Mg, and Al. The main advantages of Raman spectroscopy are noninvasiveness, reliability, and sensitivity, whereas the main disadvantages are fluorescence emission, which renders difficult the acquisition of Raman spectra, and interference of external light. In the last decades, laser-induced breakdown spectroscopy (LIBS) has been increasingly employed in studies of historical and archaeological objects, monuments, and artworks [2–6]. Repeated laser pulses were shown to be able to remove dust coatings and analyze the weathering layers, thus allowing detailed investigation of the rock varnish features and the underlying pristine rock composition [7–11]. With respect to other analytical techniques, such as scanning electron microscopy-energydispersive X-ray (SEM-EDX), secondary ion mass spectroscopy (SIMS), XRF spectroscopy, X-ray diffraction (XRD), and particle-induced X-ray emission (PIXE), LIBS offers a number of competitive features for the analysis of cultural heritage objects.

468

G. S. Senesi and O. De Pascale

These include simultaneous multielemental analysis under atmospheric conditions in a very short time, microinvasiveness, simplicity of instrumentation, almost no sample preparation, and no restriction of sample shape and size [12]. Several mobile LIBS instruments have been developed for rapid in-situ measurements both indoors at museums and laboratories and outdoors at historical terrestrial and submersed sites (see Sect. 3). However, the level of efficiency of a mobile LIBS apparatus depends on several factors, such as its weight, size robustness, and degree of independence of resources (e.g., electrical power, cooling water, gases, etc.). Mobile LIBS instruments can be distinguished in “portable” and “transportable” ones based on whether the instrument is “portable” by a single person fitting in a suitcase or backpack and is battery-operated or “transportable” by more persons often with the aid of cars or vans and requires some installation work. Further, in the last decade, several commercial compact handheld LIBS instruments have been developed and become available, which can be operated while being held in the hand of the operator in the appropriate position during the relatively short measurement time [13]. Handheld LIBS feature capabilities that rival traditional benchtop instruments, which include spectral ranges from ultraviolet (UV) to near-infrared (NIR), argon purging, rastering, and portability, all enabling field applications that were not possible heretofore. The aim of this chapter is briefly survey the results obtained in-situ and in the laboratory specifically by the use of mobile LIBS instruments, i.e., “portable,” “transportable,” and handheld, on archeological and historical cultural heritage materials, including monument stones, pigments, mural paintings, metals, and even submersed materials.

17.2

The LIBS Technique

17.2.1 Principles, Mechanisms, and Processes of Plasma Formation and Dynamics LIBS is a plasma-based atomic emission spectroscopy technique that permits rapid qualitative and quantitative multielemental analysis. It relies on ablating and evaporating a material by focusing the radiation from a pulsed laser onto the surface of the target so forming a hot plasma in which atoms and ions are excited and emit characteristic lines that can be analyzed spectroscopically. Thus, elemental information can be obtained via specific atomic or ionic transitions and associated emanating photons. The knowledge of the physics of plasma formation during laser-matter interaction and its evolution, which is a key factor to determine the optimal operational parameters of the instrument, is not yet exhaustive even if some fundamentals are known. Considering a high-power laser pulse incident onto a homogeneous target, the absorption of radiation from the beam is described by the Beer-Lambert law: I ðzÞ ¼ ð1  RÞI 0 eαz ¼ ð1  RÞI 0 e

z δopt

ð17:1Þ

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

469

where I(z) is the irradiance at distance z from the sample surface along the axis of propagation, R is the reflectivity (the factor 1  R describes the part of irradiance that enters the sample), I0 is the irradiance at the sample surface, α ¼ 1/δopt is the absorption coefficient, and δopt is the optical penetration depth. Both R and α depend on the laser wavelength via dependence on the complex index of refraction of the material [12]. If the kinetic energy produced by the beam is higher than the potential crater depth, then particles can escape. Here the potential crater depth is referred to the amount of energy needed by the structure to hold onto a single particle (similar to bond strength). Ablation (removal) of sample constituents is an extremely fast process due to the impact pressure of the laser beam that does not depend on any thermal effects and is able to extract atoms and ions from the bulk material into a vapor plume above the sample surface which travels at supersonic speed along the axis of propagation [12]. A plasma is a hot gas characterized by strong Coulomb (or electrostatic) interactions, where ionization strips electrons from atoms and positive and negative charges are balanced in a quasi-neutral state. To obtain the breakdown, some initial free electrons in the beam path must be available to absorb energy from photons, and then rapid ionization occurs [12, 14]. The high-temperature plume induces thermionic emissions that generate free electrons above the surface; thus the thermal energy overcomes the work function of the material and plasma formation is initiated [14]. The free electrons absorb energy from collisions with photons through inverse bremsstrahlung (IB) processes, and some neutral atoms can be ionized by photoionization and/or multiphoton ionization that can occur at high irradiances, so producing m photons and more electrons: A þ mhv ! Aþ þ e

ð17:2Þ

As the electron energies grow by absorption of beam energy, collisions with neutral atoms produce ionization creating more free electrons that absorb more energy from photons and ionize more atoms, and so on, initiating plasma formation [14]: e þ A ! 2e þ Aþ

ð17:3Þ

The change in electron density occurring during these processes is governed by the following equation: dne ¼ ne ðvi  va  vr Þ þ W m I m n þ ∇ðD∇ne Þ dt

ð17:4Þ

where Wm is the multiphoton ionization rate coefficient, D is the diffusion coefficient for electrons, Im is the irradiance necessary for a m-photon process to occur, n is the number density of species irradiated, and vi, va, and vr are, respectively, the rates of impact ionization, attachment (combinations of electrons with neutral atoms to form negative ions), and recombination [12]. This cascade process is called “avalanche

470

G. S. Senesi and O. De Pascale

ionization” and produces an exponential increase in the number of free electrons. Once the density of free electrons exceeds a specific value, which depends on the material and beam factors, a hot plasma visible to the naked eye is formed. The typical energy deposition needed for ablation and breakdown, i.e., the initiation of the laser-induced plasma process, is of the order of 1 GW/cm 2 [12, 14], although this value may vary depending on the target material, its physical state, and the surrounding air pressure. However, the evolution of the plasma is similar for many target types [14]. The initial period is characterized by a bright continuum emission caused by the bremsstrahlung effect, i.e., the free-free transitions of electrons occurring when photons are emitted from electron-ion interactions. This is followed by the expansion of the plasma plume along the direction of beam propagation which slows down due to collisions with the surrounding gas. When the plasma has cooled adequately, the continuum emission decreases, and the radiative recombination of ions to bound states, i.e., free-bound transitions of electrons, dominates [12, 14], and characteristic lines are emitted spontaneously due to the decay of highly excited species from upper energy to lower energy levels. Some ions and electrons then recombine to form neutral species, and some of these recombine to form molecules, the detection of which becomes possible at later times [12]. If light is collected during the entire process, the spectra can be dominated by the continuum; thus LIBS measurements are usually gated for detection after the continuum emission decays. The light emitted from the plasma is collected by a detection system consisting of a spectrometer and a camera where the light is split into its constituent wavelengths and recorded. As each element emits at specific wavelengths, the spectrum represents the plot of collected light intensities at specific wavelengths and allows the identification of the species providing the chemical fingerprint of the material [12]. Assuming that the plasma is in the local thermodynamic equilibrium (LTE) state, the density of energetic levels can be represented by the Boltzmann distribution: n ¼ no ðg=ZÞ exp ½Ei ¼ kBT 

ð17:5Þ

where * denotes an excited state and o a ground state, and the spectral line radiant intensity is given by: I mp /  ðm ! pÞ ¼ ¼


 hvmp gAmp o n exp E=kB T 4πZ

ð17:6Þ


 hcgAmp o n exp E=kB T 4πλZ

where Z is the partition function, g the degeneracy of the upper energy level, and kB the Boltzmann constant. Equation (17.6) denotes that the quantities obtained from a measurement of the plasma light are the population densities of the atomic species [14].

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

471

The LTE state of the plasma can be determined considering that the intensities of two lines of the same species with similar upper-level energies follow Eq. (17.6). Specifically, the ratio of two lines of the same species is: 0 0
 I 0 hcg Amp 4πλ0 Z no exp ðE0 EÞ=kB T ¼ I 4πλZ hcgAmp no

ð17:7Þ

0 0 0
 I 0 g Amp λ ¼ exp ðE0 EÞ=kB T I λgAmp

ð17:8Þ

If lines of known parameters are chosen, T can be calculated from the relative intensities of these lines [12, 15]. Further, the estimation of plasma temperature can be improved using several pairs of measurements. Equation (17.6) can be rearranged in the logarithmic form of a straight line with slope of 1/kBT:
 4πλZI ¼ exp E=kB T hcno gAmp  ln 

4πλZI hcno gAmp

Iλ ln gAmp



 ¼

¼

E kB T

  E 4πZ  ln o hcn kB T

ð17:9Þ

ð17:10Þ

ð17:11Þ

If the plot of ln(I λ/gAmp) versus E* for multiple species (Boltzmann plot) [15] is linear, the Boltzmann distribution and the LTE condition can be considered fulfilled, and the temperature can be determined from the slope.

17.2.2 LIBS Instrumentation A conventional LIBS equipment consists of a laser, a spectrometer, a number of lenses and optical fibers, a data acquisition system, and a control and synchronization system between the laser and the spectrometer (Fig. 17.1). The most commonly used laser is the Neodymium-doped yttrium aluminum garnet (Nd:YAG) solid-state laser that emits radiation at the characteristic wavelength of 1064 nm with a pulse of approximately 8-ns duration. The Nd:YAG laser may also be coupled to frequency doublers that allow it to operate at lower wavelengths, such as 532 nm, 355 nm, and 266 nm. Other types of laser are also used, such as the excimer laser, which emits radiation in the visible and ultraviolet region. Recently, many analysts use a Ti: Saffire laser, which allows the generation of pulses of the order of femtoseconds (fs). The high-power laser beam (~1010–1012 W/cm2) is focused on the sample causing the ablation of a small portion of matter (of the order of nanograms, ng), which

472

G. S. Senesi and O. De Pascale

Fig. 17.1 (a) Experimental scheme of a LIBS system. (b) Sketch of handheld LIBS instrument

generates a plasma plume at a temperature above 50,000 K. At this temperature, the material dissociates into ions and excited atoms emitting a continuous radiation that is not useful for the characterization of the sample. However, due to the high velocity of the species in the plasma, supersonic and adiabatic expansions occur which cool the plasma to temperatures between 5000 K and 15,000 K in less than 1 μs. At these temperatures, the elements present in the sample produce unique spectral atomic and ionic emission lines that allow their univocal identification and quantification [12, 14]. The spectrometer may be a conventional monochromator, such as the CzernyTuner that selects only a very narrow band of the spectrum and requires a sweep to obtain the entire spectrum, or, more often, a polychromator, such as the Echelle arrangement that allows the simultaneous acquisition of the whole spectrum. For the detection of the light emitted by the plasma, a charge-coupled device (CCD) detector or an amplified, intensified (ICCD) detector is used, which allows to measure the evolution of the light emitted by the plasma in steps of the order of ns with acquisition windows of some tens of ns. In particular, the ICCD detector is used when working with the plasma at a specific temperature that depends on the time of expansion, thus favoring the emission of a certain spectral line in detriment of another. The spectral range of commercial spectrometers is commonly between 190 and 900 nm, which allows to detect almost any element of the periodic table, thus representing a robust tool for materials identification. LIBS has many advantages over other methods, including instrument simplicity, rapid and simultaneous analysis of multiple elements, dust or surface removal, depth profiling, high spatial resolution (depending on laser spot size), low power consumption, close-up and stand-off capabilities, and, most importantly, no sample preparation required. LIBS is also robust with no limitation of the species detectable, and measurements can be acquired from a target in any form (i.e., solid, liquid, or gas). Although LIBS is often considered to be less sensitive than other spectroscopic emission analytical techniques, the search to improve its sensitivity and detection

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

473

limits has progressed markedly in the last decade. In particular, one of the most studied systems in recent years has been the double pulse (DP)-LIBS. In this arrangement a second laser pulse is triggered after a delay time (interpulse) from the first laser pulse by re-exciting either the region where the plasma is generated by the first pulse or, in some cases, the surface of the sample [12, 14]. A great interest has been and is devoted to the miniaturization of LIBS system components in order to develop portable equipment able to perform in situ analysis with increased stability and sensitivity of the measurements [13, 16, 17]. An example of the maturity achieved by LIBS is provided by its adoption in space missions by important spatial agencies, such as the National Aeronautics and Space Administration (NASA) and the European Space Agency (ESA). The most recent case of LIBS use in the space is represented by the Curiosity rover equipped with a LIBS instrument used in the Mars space mission to analyze Martian soil [18].

17.2.3 Quantitative Analysis Besides the identification of elements on their basis of the emission wavelengths, LIBS spectra can provide quantitative information on the element concentration in the sample if a predictive model, i.e., a calibration curve, is constructed by plotting the peak intensities versus the known concentrations of standard materials. In particular, the calibration curve for a specific element in a material should feature a linear relationship between the integrated intensity Ij of a representative element line at a specific spectral wavelength λj and the known concentration of the element. The quantitative LIBS analysis of complex matrices, e.g., archeological and historical materials, can be, however, hampered by the so-called matrix effects on the intensity of emission lines, which are related to differences in material matrices, including hardness, chemical composition, density, and reflectivity, and can result also from shot-to-shot variations in the vaporized mass and/or in the excitation temperature of the plasma which, in turn, result from fluctuations in laser intensity, coupling efficiency of the laser with the target material, and/or other instrumental variations [12]. Although the dynamics of matrix effects have been studied widely also recently, more in-depth studies are required to eliminate or at least minimize them. Because of the difficulty in correcting the matrix effects in the past, LIBS could not be used for elemental quantification in the early times. However, the advancement of electronics, laser stabilization, and study of plasma dynamics have changed the scenario rapidly and made LIBS adequate for quantitative analysis so gaining a renewed attention worldwide. A common approach to correct matrix effects in LIBS measurements is the use of an internal standard for normalization of peak intensities [12, 14]. In this case, calibration curves are constructed using intensity ratios Ij/Ir where the normalization intensity Ir is usually that of the line of the major element present in the matrix preferably of all samples [14]. Standards that have a varying amount of the relevant element are used, then the spectrum is measured from each sample, and the intensity ratios at each concentration are plotted. The efficiency of a calibration curve is evaluated by its linear regression analysis coefficient [14].

474

G. S. Senesi and O. De Pascale

In a calibration curve, each point along the fitted line is the integrated peak intensity from different samples containing the element of interest J. For an unknown sample, the integrated intensity of the peak of the element J is measured, and its concentration is indicated on the abscissa of the calibration curve [14]. A specific calibration curve must be constructed for each element detected in the spectrum. An alternative to calibration approaches is the so-called calibration-free (CF) or self-calibration method, which is based on the LTE assumption and was first proposed by Ciucci et al. [19]. In particular, the LTE condition implies that the number density of emitters, Nu, is given by the Boltzmann distribution: Nu ¼

  gu N 0 E exp  u kT Z ðT Þ

ð17:12Þ

where gu is the statistical weight of the excited level of energy Eu from which the emission occurs, N0 is the total number density of the given species, and Z(T) is the partition function at the excitation temperature T. By substituting Nu in Eq. (17.12) and putting it in a logarithmic form, the equation of a line is obtained: ln

I ul 4π N G E ¼ ln 0  u Aul hvul gu Z kT

ð17:13Þ

where an instrumental factor, G, appears which cannot be evaluated unless a radiometric calibration is performed. This method, commonly referred to as the Boltzmann plot method, enables the determination of the excitation temperature, T, from the line slope and N0 from the intercept. N0 is proportional to the weight percentage of the species i in the sample and can be retrieved as described in [19] by a normalization procedure: X X w%i ¼ x N 0i AW i ¼ 1 i

ð17:14Þ

i

where AWi is the atomic weight of the species i and the normalization constant x also contains the instrumental factor G. Once x has been determined, each species percentage is given by: w%i ¼ x N 0i AW i

ð17:15Þ

The CF algorithm represents the only possible approach when the preparation of matrix-matched standards is unfeasible or difficult, e.g., in extreme environments, such as nuclear reactors, and in automated analytical procedures. However, in order the CF approach can be used for quantitative measurements, the following assumptions must be strictly fulfilled: stoichiometric ablation, LTE condition, plasma homogeneity, and optical thin plasma. Thus, the accuracy of CF quantitative measurements depends critically on these assumptions, and it may occur that only semiquantitative results can be obtained when their fulfillment is unsatisfactory.

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

17.3

475

Applications of Portable and Handheld LIBS Instrumentations to Cultural Heritage Materials

17.3.1 Stone Materials Weathering and alteration of stones are a result of a number of various physical and chemical processes occurring between the stone surface and the surrounding environment. A correct diagnosis of the deterioration state at both the surface and the bulk of the material is very important for choosing the most appropriate means for restoration and conservation interventions. In particular, mineralogical, petrographical, textural, microstructural, and chemical analyses, especially for the identification and concentration gradient of some contaminant elements, are needed to be performed along the depth profile of the material under study. Traditionally, this analysis requires sampling of the object and subsequent laboratory measurements by diagnostic techniques that in the past have often caused serious problems to the artifacts examined. In the last decade, mobile (portable and transportable) LIBS instruments have shown a promising performance in studying the weathering and alteration processes of various stone (and other) materials caused by environmental factors. In particular, LIBS can perform successfully a micro-destructive, in-situ, multilayer diagnostic analysis and in-depth elemental profiling of encrustations and impurities even for highly inhomogeneous layered crusts before proceeding to conservation treatments. Repetitive laser shots at the same spot can ablate a thin weathered (~100 μm thick) layer of the material so recording changes in composition as the underlying bulk material is approached. A detailed, in-depth profiling LIBS study was conducted in situ by MaravelakiKalaitzakia et al. [8] on encrustations covering various exposed Pentelic marbles from ancient Greek monuments. In particular, the samples studied were the dendritic and thin black encrustations on the marble walls of the Adrian Bibliotek in the Foro Romano in Athens, the soil-dust depositions on excavated artworks, and the patina samples collected from Roman walls in the same area. LIBS performance was compared to that of other analytical techniques requiring removal of sample from the monument, such as optical microscopy (OM), XRD, Fourier transform infrared (FTIR) spectroscopy, and SEM-EDS. The analysis of individual LIBS emission spectra measured upon irradiation of different types of samples with successive laser pulses provided the stratigraphic trends of the relative concentrations of the elements present. The elemental LIBS profiles of the dendritic black encrustations based on the relative spectral line intensity values showed that the concentration of Fe, Si, Al, and Ti, relative to that of Ca, decreased significantly with depth (Fig. 17.2), thus indicating a decreasing contamination and deposition originated from atmospheric pollutants down to the alteration layers. In particular, two stratified alteration layers were identified by both SEM-EDX and LIBS in the black and compact thin crust usually present as a continuous, firmly attached layer on the outer surface of marble in areas of the monument which were subject to intense percolating rainwater, i.e., underneath cornices and pillar crowns. The external layer featured the

476

G. S. Senesi and O. De Pascale

Fig. 17.2 The LIBS spectra of the ablated material from a dendritic crust at different depths from the surface and unweathered Pentelic marble, normalized to the intensity of Ca II line at 317.93 nm. (Reprinted from [8], with permission from Elsevier)

maximum intensity of Fe, Si, Al, and Ti lines that showed by an intense decreasing trend with depth and consisted of Al silicates, gypsum, and particulates, whereas the internal layer appeared almost homogeneous in composition. In soil-dust deposits, both techniques identified an external alteration layer featuring decreasing gradients of Fe, Si, Al, S, and Ti and consisting of gypsum and aluminosilicates and an internal layer characterized by the presence of Si and absence of S and Ti and constituted of only Al silicates. In the remnants of patina layer samples from original treatments made for protective purposes, both techniques revealed rather stable in-depth values of Si and Al, a low concentration of S, and absence of Fe and Ti. Although some differences were found between LIBS and SEM-EDX analytical results for the external encrustation layers, possibly due to their inhomogeneity and roughness that affect laser-matter coupling, LIBS appeared to be an autonomous, rapid, insitu, micro-destructive diagnostic mean to obtain in-depth elemental profiling of layered crusts on Pentelic marbles. Grönlund et al. [20] performed successfully a number of remote imaging and remote ablative cleaning LIBS experiments on some cultural heritage objects by using a fully self-contained, mobile, lidar system comprising a tripled Nd:YAG laser working at 355 nm with 170 mJ pulse energy with an expanded beam focused onto the target at 60-m distance. The LIBS signal was detected by using an on-axis Newtonian telescope and an optical multichannel analyzer, whereas the imaging was performed by scanning the laser beam on the target. Imaging processing of the

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

477

characteristic lines of different elements allowed to identify the target material. The same setup was also used successfully for remote laser ablative removal of biodeteriorated surface layers on a granite artifact, i.e., areas heavily blackened by soft pencil and marker pen on Volterra white alabaster slabs and an Italian replica statue head made of white Carrara marble, and also for cleaning an Italian replica statue partially covered by dirt and algae. Thus, remote imaging and remote cleaning LIBS appeared to be a viable technique for surface operations on historical monuments at considerable stand-off distance. Fortes et al. [21] used a man-portable LIBS instrument for the in-situ, real-time spectral acquisition and chemical mapping of the façade of the cathedral of Malaga, Spain, in order to discriminate qualitatively among sandstone, limestone, marble, and cement mortar, i.e., the main components used in this historical building. A protocol of analysis was chosen so to achieve an accurate description of the building materials with adequate spatial resolutions. The apparatus consisted of three distinct pieces, i.e., a compact handheld probe, a main unit, and a laser power supply which were interconnected by cables. In particular, a Q-switched Nd:YAG laser operating at 1064 nm was utilized to generate the LIBS plasma on the sample surface, the spectrochemical analysis of which provided the chemical characterization of the different materials employed in the construction of the building. Significant areas of the cathedral, i.e., six zones of the northern façade and the “girola,” were analyzed in the field by the portable apparatus. Results obtained by analyzing the spectral window covering the range 240–360 nm, where the most important emission lines of the elements of interest are present, confirmed those achieved in a systematic LIBS study conducted in the laboratory. In particular, (a) the zones whose spectra showed an emission peak of Si I at 288.16 nm much more intense than that of Ca I at 315.88 nm were constituted of sandstones; (b) the zones made of limestone featured high intensity of Ca II and weak emissions of Si I and Mg I at 285.21; (c) marble zones were identified by strong emission lines of C, Ca, and Mg; and (d) the presence of cement mortar was characterized by spectra similar to those of limestone but with emission intensity of Mg higher than that of Si. The heterogeneity and morphology of the northern portal stones, which were evaluated by measuring the net intensity of C, Ca, Mg, and Si in two typical depth profiles, showed that stones of this area were quite clean and free from typical encrustations due to salts and pollution deposits. Further, large-area compositional maps were constructed using the portable LIBS system to obtain the spatial distribution of the main surface constituents of the structural stones. In particular, chemical images of Si/Ca and Ca/Mg intensity ratios were generated from the northern façade and the “girola” with total areas analyzed of 250 m2 and 650 m2, respectively. Large concentrations of Si were found in the northern portal, which suggested sandstone as the structural stone in this area, whereas Ca was present at a significant extent only in marble columns and the remaining portal wall, which indicate limestone. In conclusion, the sensitivity and versatility of the analytical approach based on portable LIBS allowed to discriminate successfully between construction materials on the basis of different ratios of intensities. Further, the large-area compositional mapping generated by LIBS yielded the spatial distribution of elemental constituents of the structural stones.

478

G. S. Senesi and O. De Pascale

Finally, the LIBS system was able to provide further invaluable information on stone clearness, deterioration, conservation, and potential chemical treatments to be made on the monument surface stones. A mobile DP-LIBS instrument called Modì (Mobile Double Pulse Instrument for LIBS analysis) was used by Brai et al. [22] to evaluate the potentiality of LIBS for the rapid analysis of cultural heritage materials, in comparison to the portable micro (μ)-XRF instrument ArtTAX. The comparative analysis was performed on calcareous and refractory building fragments, i.e., bricks and well-preserved and degraded mortars, located in different areas of the ancient Greek-Roman Theatre of Taormina, Sicily, Italy. Certified materials were used as reference to construct calibration curves by XRF measurements and to evaluate LIBS analytical performance. Although LIBS data showed a nonlinear correlation to concentration, thus no quantitative results could be achieved by LIBS, and the three types of building materials analyzed could be grouped into three distinct clusters. The low accuracy of LIBS measurements, with respect to XRF ones, could be ascribed to the interactions laser-matter, i.e., to the different amount of mass ablated and plume composition, which led to errors of the element percentages per unit of mass ablated on dependence on the different material features, e.g., density, compaction, and chemical composition. The authors concluded that, although the integrated study of XRF and LIBS signals permitted to distinguish among chemical features and degradation state of the building materials analyzed, more studies are needed to improve LIBS repeatability and reliability by better understanding and correcting related problems. A LIBS instrumentation easily compacted into a movable platform was used by Gaona et al. [23] for the in-situ and stand-off spectroscopic analysis of parts of the Cathedral of Malaga, Spain, from an average distance of 35 m without any need of scaffolding or other intrusive facilities. In particular, the comprehensive characterization of the materials composing the central wing of the lower level of the main façade, including the medallion located therein and the flanking ornamental elements, and the identification of the potential pollutants spoiling their surfaces have been performed by a stand-off LIBS sensor using a high-power DP Nd:YAG laser system operating at 1064 nm. The LIBS results fitted neatly with the mineralogical analysis of all the stones assayed. In particular, the intense emissions of Si I at 390.6 nm, Al I at 396.1 nm, Ca II at 396.9 nm, and Mg I at 393.8 nm with faint emissions of Mn I at 403.1 nm, Sr I at 407.8 nm, and Fe I at 404.6 nm confirmed that the main structure was almost entirely built using sandstone. Further, the spectra of three marbles of different color (white, black, and rose) constituting the medallion located in the front façade were all dominated by atomic and ionic emissions of Ca with small contributions of Al and Ti. However, the spectral features of Mg I and Sr II in black and white marbles were more intense than in rose marble, and of comparable intensity in black marble, with more abundant Mg in white marble. The ratio of intensities Sr I to Ca I and Mg I to Ca I allowed to group the three marbles into three distinct clouds (Fig. 17.3a), thus confirming the capability of stand-off LIBS to sorting consistently materials of the same kind and identifying their provenance that resulted to be the same for marbles of the same color. Stand-off LIBS also allowed the detection of pollutants deposited in some areas on the façade.

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

479

Fig. 17.3 (a) 3D scatter plot projected onto the subspace of the Sr/Ca and Mg/Ca emission intensity ratios for series of 500 single-shot LIBS spectra gathered each from 3 types of marbles composing the most distinctive medallion and distinguishable by different colorations (white, black, and rose). Reprinted from [23], with permission from Royal Society of Chemistry. (b) Surfaceaveraged LIBS intensities of the elements considered at five laser pulses in depth. (Reprinted from [24], with permission from Elsevier)

In particular, LIBS analysis of a Solomonic column made of rose marble showed that the apparently clean downward parts featured new lines of Fe, Si, Mn, Pb, and Ba, besides the expected emissions of Ca, Mg, and Al and small contributions from Ti and Sr, which were more intense in the crust-coated upward parts. The in-depth profile of this sample and the calculated intensity ratio of several elements to the Ca I signal confirmed the above findings. In particular, the presence of Ba and Sr could be related to environmental sources, i.e., the deposition of resuspended dust from the Sahara desert and atmospheric particulate matter such as marine aerosols from the sea coast located nearby. Differently, trace elements such as Ti, Pb, and Mn would originate from anthropogenic sources, i.e., soot and metallic particles bearing Pb from exhausts of gasoline and diesel engines. Further, the analysis of three colored marbles from one of the corners of the main façade allowed to line up the three corresponding datasets in three distinct clouds, thus indicating that the pollution extent changed from one area to another. In conclusion, all the information obtained by the stand-off LIBS system were of primary importance prior of planning any restoration program. Recently, Senesi et al. [24, 25] investigated by two different mobile LIBS instruments the elemental composition of the black crust deposit accumulated over the centuries on the surface of a limestone fragment collected from a quoin of the left jamb of the southern entrance gate to the courtyard of Castello Svevo, Bari, Italy, which was previously analyzed by a common benchtop LIBS [10, 11]. In the first work, Senesi et al. [24] used a mobile μLIBS Modì instrument working with a DP Nd:YAG laser emitting two collinear pulses at 1064 nm in conjunction with OM. The laser beam was directed through the Modì articulated arm to the optical microscope and then focused on the sample. The LIBS signal was collected by an

480

G. S. Senesi and O. De Pascale

optical fiber and analyzed by a double spectrometer covering the spectral range from 190 to 900 nm. The surface-averaged elemental maps of LIBS spectra indicated that the concentration of the main elements in the black crust overlying the limestone, i.e., Al, Fe, Si, Ti, Na, Sr, and Mg, decreased with increasing depth, except C and Ca, i.e., the main components of the limestone (Fig. 17.3b). The presence of Al, Fe, Si, and Ti in the black crust was ascribed to atmospheric pollution, dust, and particulate deposition, whereas that of Na might be due to the marine aerosol that pervades the urban area where the monument is located. The decrease of Sr with depth also suggested an environmental source, whereas that of Mg indicated a possible dedolomitization of the stone. The three-dimensional (3D) compositional maps generated from elemental imaging performed on the sample rough surfaces without any preparation confirmed the above trends. Further, the DP-μLIBS analysis of sliced “virtual thin sections” (VTS) processed by the self-organized maps (SOM) approach allowed to better estimate the extent of the alteration processes that occurred at the limestone surface and the change of the mineralogical composition in the transition from the black crust to the limestone underneath. Thus, the mobile DP-μLIBS featured a high potential to perform a rapid and detailed in-depth compositional and mineralogical analysis, including elements of low atomic mass, such as C, of weathered artifacts by using high-resolution and fast 3D compositional imaging and VTS. In the other work [25], the performance of a self-contained, portable, handheld LIBS instrument was tested for the first time for the direct, real-time, elemental analysis in atmospheric air to compare and discriminate between two different portions, i.e., the weathered layer and the underlying stone surface, of the same monument sample analyzed by the Modì LIBS instrument [24]. The handheld LIBS instrument used named NanoLIBS consisted of a miniature-diode-pumped, solidstate, short-pulsed laser emitting at the wavelength of 1064 nm and equipped with a compact spectrometer covering the spectral range from 180 to 800 nm. The whole setup was enveloped in a lightweight handheld body with mass and dimensions of about 1.8 kg and 26  10  30 cm, respectively. The measurements were performed by placing the nose of the instrument against the sample and then starting the analysis via a trigger. The instrument was provided with a rechargeable Li-ion battery allowing up to 8-h operation. The “chemical fingerprint” of the superficial deteriorated stone sample was provided by full broadband emission spectra that, when examined in detail in specific spectral ranges (Fig. 17.4), clearly indicated the presence of Al, Fe, Si, Na, and K as the contaminant elements in the black crust, whose origin was suggested previously [10, 11, 24]. These elements were not detected in the underlying unaltered limestone where the expected presence of the elements Ca, C, and Mg was confirmed. The comparison of quantitative data obtained by the matrix-independent CF approach using the handheld instrument with those obtained using the benchtop system showed differences lower than 25% for all the elements, except for C and Ca that yielded a difference of about 50%. Thus, the commercially available, easily operated handheld LIBS instrument showed a good performance for the rapid, direct, in-situ elemental identification and discrimination of weathered artifacts without requiring any removal and

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

481

Fig. 17.4 LIBS spectra of the black crust and limestone at four specific spectral ranges, i.e., from 230 to 270 nm (a), from 270 to 310 nm (b), from 575 to 595 nm (c), and from 710 to 780 nm (d). (Reprinted from [25], with permission from Elsevier)

preparation of the sample. Although only semiquantitative data could be obtained by the handheld instrument, in the future the use of appropriate multivariate statistical LIBS data processing is expected able to unveil the classification and provenance of the raw parent materials, thus providing archeologists, historians, and conservators unique information on the historical context and chronological period to which the object studied belongs. Recently, Aramendia et al. [26] used a novel approach, i.e., a combination of Raman spectroscopy and LIBS, to determine in-situ the concentration of hydrated mineral phases by estimating the H2O % by LIBS and that of non-hydrated compounds by Raman in a number of calcitic and dolomitic marble samples with or without a patina of calcium oxalate (whewellite) originating from various historical buildings in Spain, Italy, and Serbia. In particular, a portable EasyLIBS instrument operating with a Nd:YAG laser at 1064 nm and equipped with three spectrometers covering the UV, Vis, and near-IR spectral ranges was used in conjunction with a handheld InnoRaman spectrometer provided with a 532-nm excitation laser and a 5-m-long microprobe. LIBS analysis allowed to obtain the H2O % in marble samples using a calibration curve constructed using standard samples of different H2O content. In particular, to estimate the H2O %, the area of the main LIBS H band at

482

G. S. Senesi and O. De Pascale

656.28 nm present in all standards was plotted versus the known H2O %. As the LIBS H band was overlapped by secondary bands of Ca at 657.46 nm, C at 657.98 nm, and O at 655.20 nm, a decomposition of the H band was performed for the LIBS spectra of standards and samples. Once constructed the calibration curve, to obtain comparative results, the marble samples with or without the calcium oxalate patina were measured by LIBS in the same conditions, i.e., the same acquisition parameters as the standards and the Raman measurements, that is, the same size and location of spot areas and equally homogeneous spots were analyzed. The oxalate % obtained by both benchtop and portable point-to-point Raman and portable LIBS resulted comparable in the areas where the oxalate patina was present, which showed that both techniques were appropriate to discriminate mineral phases containing crystallization H2O in marble samples. Further, large differences were measured in whewellite concentrations on the surface analyzed, i.e., from nearly zero on calcitic areas up to 30% in dolomitic ones. In conclusion, the combination of molecular and elemental cross information obtained, respectively, by Raman and LIBS appeared to overcome the incomplete and sometimes misleading information obtained by using Raman only and crucial to achieve complete and reliable quantitative results. Further, the approach used in this work appeared very promising for application to any kind of stone samples presenting surface efflorescences.

17.3.2 Pigments and Mural Paintings The characterization and identification of pigments used in paintings and in other various heritage objects and monuments are of high artistic, technical, or historical relevance. In particular, the analysis of pigments and colorants can provide invaluable information to archaeologists, historians, and conservation scientists on painting techniques used and their evolution across the centuries. Further, these information are expected to be relevant in supporting archaeometric studies related to provenance, dating, and/or authenticity, detecting material degradation and its causes and mechanisms, developing proper methods of conservation and restoration, and evaluating their effectiveness. In a very exhaustive study, Westlake et al. [27] tested a multi-analytical approach, which combined mobile LIBS and micro-Raman instrumentations complemented by laboratory techniques, i.e., XRD, OM, and SEM-EDX coupled to microprobe analysis and X-ray mapping, to explore directly on the wall on site 49 Cretan wall painting pigment samples spanning two millennia of history with the aim of unveiling the evolution of painting materials and technology from the Bronze Age (Minoan) to Roman-Hellenistic and Byzantine periods. In particular, painting samples with little or no evidence of modern conservation interventions from the Bronze Age palace and two Roman buildings at Knossos and the Byzantine Metropolis Basilica at Gortyna, Crete, were analyzed. The portable LIBS instrument used consisted of a small-sized Q-switched Nd:YAG laser and a compact fiber optic spectrometer covering the spectral range from 240 to 760 nm equipped with a miniature CCD camera to visualize the object placed at about 70 mm from the

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

483

optical head during analysis. The mobile Raman microspectrometer used a continuous wave (CW) diode laser emitting at 786 nm as the excitation source, which was fiber optically coupled to an optical head that allowed variable focusing of the beam on the sample surface. In the case of polychromatic samples, several spots across the painted surface were analyzed in situ by portable LIBS and Raman instruments, whereas a smaller but representative number of fragments selected on the basis of results were analyzed for confirmation by XRD and SEM-EDX laboratory instruments. In particular, LIBS spectra collected on different tonalities of red and pink color areas of wall painting fragments showed typical emission lines of Fe of various intensities correlated to the density of the Fe-based pigment used, which most likely were red ochre consisting mainly of Fe oxides/hydroxides. Additional elements detected were Ca, Na, Al, Mg, Si, trace amounts of Ba and Sr, and occasionally Zn and Pb, all originating from matrix minerals such as calcite, clay minerals, and quartz. LIBS analysis of yellow paints also showed intense emissions from Fe originating from yellow ochre, i.e., goethite and/or limonite, and additional matrix emissions from Al, Ba, Ca, Mg, Na, Si, and Sr. Raman and XRD analyses confirmed LIBS results proving the presence of hematite/iron ochre and accessory minerals in red paints and yellow ochre/goethite and limonite in yellow paints. Blue paints featured LIBS spectra dominated by strong Ca emission lines with emission from Cu that resulted relatively weak due to red or purple paint applied over the underlying blue layer, but clear enough to indicate the use of a Cu-based blue pigment that could not be further identified by LIBS data. The same additional matrix emission lines as in the yellow paint were also detected. The green color paints were present only in paintings of the Roman and Byzantine periods, and their LIBS analysis yielded an elemental profile featuring Al, Ba, Ca, Fe, Mg, Na, Si, and Sr, failing to detect Cu in most samples. However, both Raman and XRD data enabled the identification of the Cu-based pigments used for the blue color as cuprorivaite, the so-called Egyptian blue, which was mixed with yellow ochre for generating green. Strong LIBS emissions from Al, Ca, Na, and Si and a weak emission from Fe were detected on the gray-to-black paintings, which pointed to the use of a C-based black pigment possibly originating from burnt vegetal materials. LIBS analysis of white color areas, which were present only in a few fragments, indicated Ca as the dominant element, which suggested a lime-based pigment. In general, the wall painting palette used in Crete from the Bronze Age to the early Byzantine period appeared quite limited, i.e., few changes could be observed in the materials and techniques used. As a whole, the LIBS system used appeared to be more versatile than the Raman system as it regarded operational factors, including the capacity to analyze large objects or structures, e.g., a wall, by using a scaffolding, sample positioning, signal intensity and hence sensitivity, immunity to ambient light, and speed of data acquisition. However, key advantages were offered by both LIBS and Raman, which consisted in instrument mobility allowing the quick surveying of wall paintings and rapidity of data collection and interpretation on site, which permitted a targeted sample selection for further analysis by additional laboratory techniques. In conclusion, the various techniques used offered complementary information, i.e., elemental compositional data from LIBS, molecular data from Raman, and crystal

484

G. S. Senesi and O. De Pascale

structural information from XRD and SEM-EDX, which are all of great interest to archeologists and conservators. An oil painting on canvas of unknown time and origin located in Gliwice, Poland, was studied by Kaszewska et al. [28] with the aim of a planned restoration using a mobile LIBS system provided with a Q-switched Nd:YAG laser operating at 266 nm. A preliminary investigation conducted by optical coherence tomography (OCT) revealed a complicated, multilayered structure of the painting, undocumented past restoration, possible overpainting, material alteration, the presence of gaps and cracks on the surface, possible delamination, and other possible manipulations. The inspection by OCT cross-sectional images was thus necessary to exclude from LIBS analysis areas where laser firing could inflict serious damage and properly select for analysis areas of the painting that could avoid misinterpretations and ensure the representativeness of LIBS results. The depth-profile LIBS and OCT stratigraphic analysis of seven most interesting spots on the painting allowed to recognize four layers of various thickness for which the pigments were identified to assist in its dating. The varnish first layer, the semitransparent greenish blue second layer, and the blue-painted third layer showed LIBS signals from Fe, CN, and Ca which decreased in intensity with depth and Al and Mg signals of constant intensity. Further, Si and Na signals resulted more intense in the third layer, whereas Mn was present in the fourth red ground layer. In a more detailed inspection, the first layer was excluded from further investigation because it was suspected of possible surface contamination and apparent lack of pigments. The high concentration of C and N and lack of Si in the second layer pointed to Prussian blue pigment and excluded ultramarine as responsible of the bluish tint, whereas the third layer was rich in all components of ultramarine pigment. The high signal intensity of Fe, Al, Si, and Mn in the fourth red ground layer pointed to a pigment mixture based on Siena red and manganese brown, whereas the very strong lines from Pb in all layers, including dark areas, suggested an admixture of lead white pigment in all paints. In conclusion, the mobile combined setup proposed in this work integrating a high spectral resolution LIBS system with a high spatial resolution spectral OCT instrument was able to enhance significantly the quality and accuracy of the in-depth stratigraphic analysis of scaled concentration profiles of elements contained in the pigments used in each layer of the painting. Further, OCT mapping employed prior to LIBS analysis assisted in the selection of specific areas of interest on the painting surface to be examined in more detail. Papliaka et al. [29] employed jointly two distinct, complementary, mobile, and compact portable laser-based units, i.e., a LIBS system using a compact passively Q-switched cNd:YAG laser emitting pulses at 1064 nm and a Raman microspectrometer featuring an excitation diode laser working at 786-nm fiber optically coupled to an optical head, to conduct two campaigns aiming to monitoring in detail the pigments used to decorate various stone sculptures in Crete, Greece. The indoor campaign focused on two Venetian incised inscriptions (thirteenth to seventeenth centuries), five Ottoman relief inscriptions (seventeenth to nineteenth centuries), and one marble fragment difficult to assign to a certain historical context, which were hosted in a museum and a conservation laboratory at Heraklion. The outdoor

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

485

campaign aimed to analyze the iconographic and decorative paints on stone carved door frames dated to the fifteenth century and located in a monastery and a church in the rural area around Heraklion. In general, results of LIBS analysis were either complementary, confirming molecular Raman results, or unique in providing the elemental composition information. In particular, LIBS spectra of one Venetian stone inscription featured a strong molecular emission from CN indicating the presence of materials containing C, in agreement with Raman analysis. The other Venetian inscription featured LIBS emissions of Mn, which suggested the presence of Mn-based black pigment, less intense lines of Pb and Cu suggesting a mixture of a Cu-based blue pigment and lead white, and Mg and Sr lines most likely related to the carbonate matrix. LIBS analysis of the six Ottoman printed relief inscriptions provided different results for each of them, including the presence of (a) Pb ascribed to red lead pigment; (b) Hg derived from cinnabar and/or vermillon pigments; (c) Pb and Cr due to chrome yellow pigments; (d) Ca and Ba in white-painted areas suggesting the presence of lime and/or Ba sulfate; (e) Cu ascribed to blue and/or green pigments such as ultramarine and Prussian blue; (f) Pb and Cr in light green areas; (g) traces of Ag, Au, and Cu in three samples ascribed to an Au-Ag-Cu alloy used as a decorative coating (Fig. 17.5a); and (h) lines of Al, Mg, Fe, Ti, Si, and Sr detected in all samples and originating from the Ca carbonate and sulfate matrix. The scattered traces of paint on the surface of the marble fragment featured Pb, CN, and Cu LIBS lines (Fig. 17.5b) suggesting the use of red, dark, and dark-green pigments, i.e., red lead, carbon black, and malachite or verdigris. The LIBS analysis of the mantle of Virgin Mary painted on the tympanon over the door and on the architrave of Valsamonero Monastery showed, respectively, lines of Fe and Mn, most likely ascribed to umber pigment (Fig. 17.6a), and Fe but not Mn suggesting a red Fe oxide pigment. The black areas of the paintings featured strong CN emission (Fig. 17.6b)

Fig. 17.5 (a) LIBS spectrum from remains of a decorative gold foil on the Ottoman stone inscription AI244 implying the presence of a gold-silver-copper alloy. (b) Emission lines corresponding to Cu in the LIBS spectrum obtained from a green-colored area on the marble fragment from the St. Catherine collection. (Reprinted from [29], with permission from Springer)

486

G. S. Senesi and O. De Pascale

Fig. 17.6 LIBS spectra collected during the campaign at the Valsamonero Monastery (a) from the dark red mantle of Virgin Mary, suggesting the presence of umber, and (b) from a dark-painted area with the strong CN emission indicating the presence of a carbon-based paint. (Reprinted from [29], with permission from Springer)

implying the use of a C black pigment, whereas Ca was dominant in all white areas suggesting CaCO3 or CaSO4 pigments. The elements Al, Mg, and Ti were also detected by LIBS and attributed to matrix minerals such as calcite, gypsum, and/or kaolinite. The LIBS analysis of paints in the church of St. George indicated the presence of (a) Fe and Mn in red and brown paints, which suggested a mixture of Mn black or brown pigment mixed with red Fe oxide; (b) Cr, Pb, and Fe in yellowpainted areas suggesting a mixture of Cr yellow and yellow ochre pigments; (c) Fe in light blue areas possibly deriving from Prussian blue mixed with ultramarine blue; (d) strong lines of Ba and Zn detected in all paints, suggesting the presence of BaSO4 and ZnO; and (e) Al, Mg, Sr, and Ti from matrix minerals detected all across the door frame. In conclusion, the proposed analytical methodology confirmed the importance of mobile LIBS as a versatile technique complementary to mobile Raman to provide a very detailed elemental and molecular characterization of pigments and minerals, superior to the one that could be obtained by using individually LIBS and Raman datasets, and able to satisfy the demands for immovable objects routinary analysis in museums and conservatory laboratories and at historical and archeological sites. Siozos et al. [30] developed a novel hybrid portable instrument to implement an existing portable LIBS spectrometer used previously [27] with a diffuse reflectance spectroscopy (DRS) module which was tested on two sets of original pigment samples and applied to the analysis of paint on a Minoan pottery sherd. The performance of the hybrid instrument was evaluated with the diffuse reflectance module covering the entire visible spectral range 380–950 nm and the LIBS unit operating in the range of 250–660 nm. The first set comprising three Cu-based pigments, i.e., azurite (blue), malachite (green), and Egyptian blue (blue), yielded very similar LIBS spectra for azurite and malachite, which were dominated by Cu

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

487

emission lines and the presence of also Ca and Al, and a LIBS spectrum of Egyptian blue featuring the presence of Cu, Si, and Ca, with the latter two elements that might be due to the presence of CaCO3 and Si in the pigment. No conclusive results could be derived by LIBS spectra also for the second set of paints based on Fe minerals and including red Fe oxide, yellow ochre, green earth, and mars black, all presenting dominant Fe emission lines. However, DRS spectra showed distinct differences among both the three Cu pigments and the four Fe-based pigments. The LIBS spectra of the red and black slip decoration on the pottery sherd were dominated by Fe emission lines likely suggesting black and red Fe oxides, whereas the white areas were dominated by intense Ca peaks with weak Fe and Al peaks indicating a lime-based paint with impurities from the underlying clay. The DRS spectra of the three colored areas in combination with LIBS data indicated red ochre as the main red pigment mixed with yellow ochre, mars black as the black pigment, and CaCO3 or CaSO4 mixed with yellow ochre as the white pigment. In conclusion, combining complementary data obtained from the LIBS and DRS modules of the hybrid spectrometer allowed to discriminate effectively among different types of pigments even if they had the same color and/or similar elemental composition and provided more reliable information on the pigment identity if compared to using results from each technique individually. Recently, Lazic et al. [31] used three different spectroscopic techniques, i.e., portable and benchtop LIBS, portable μ-XRF and XRF scanning systems, and a benchtop PIXE to analyze various typologies of cultural heritage materials. In particular, a stand-off LIBS instrument with a Quantel Q-smart laser source operating at 1064 nm was used at the target distance of 9.5 m to analyze (i) three egg tempera pigments, i.e., cinnabar (HgS, intense red), ochre (brown-red), and a mixture of azurite and lazurite (blue), painted over a gypsum ground on wood substrate, and (ii) five oil pigments, i.e., orange red minium (Pb3O4), dark red cuprite (Cu2O), manganese brown (Mn3O4), black bone (Ca3(PO4)2), and realgar (As4S4) spread over an “imprimitura” layer of either green earth or lithopone on gypsum also placed over a wood substrate. LIBS could recognize the three egg tempera pigments remotely in a single-shot acquisition by detecting the emission lines of (i) mainly Hg, but also Mg, Na, Ca, and K in cinnabar; (ii) mainly Fe with Ca, Mg, and other minor elements in ochre, a natural earth pigment based on Fe oxides/hydroxides; and (iii) Cu (I) and C (I), originating from azurite (a Cu carbonate), and mainly Al, Si, and Na attributed to lazurite (a complex S-containing Na silicate) in the blue pigment. LIBS stratigraphic analysis showed that the emissions from Ca grew rapidly with depth reaching a maximum stable intensity, indicating the gypsum layer, after 9, 8, and 13 laser pulses, respectively, for cinnabar, ochre, and blue pigment. Further, remote LIBS depth profiling analysis provided well-resolved profiles of each of the three layers of the five oil-based pigments identifying a number of elements much larger than XRF and PIXE in the main pigment constituents in the top layer. In particular, the main elements detected were (i) C, Pb, and O in minium; (ii) Cu in cuprite; (iii) Mn and O in manganese brown; (iv) P, Ca, and O in black bone; and (v) As in realgar (Fig. 17.7). However, besides the main elements listed above, many other minor ones of various emission intensities were detected in these pigments and in the underlying

488

G. S. Senesi and O. De Pascale

Fig. 17.7 A characteristic portion of a single-shot LIBS spectrum from oil-based pigment: (a) sample 4, black bone on green earth; (b) realgar on lithopone. (Reprinted from [31], with permission from Elsevier)

second layer constituted of green earth or lithopone. In conclusion, LIBS featured a higher analytical sensitivity, chemical discrimination capacity, and in-depth profiling performance and a lower operational cost that the other techniques used, allowing the unique feature for remote in-situ analysis. However, some problems were encountered in LIBS analysis, i.e., the difficulty to detect S in general, As in the presence of Fe, and O only when in excess with respect to the air content, which could be solved by the combined field use of LIBS and XRF deserving only the most relevant samples for the expensive PIXE laboratory analysis. Also recently, Veneranda et al. [32] conducted a study aiming at identifying and comparing the conservation state of the surface of two walls of two locals located in different environmental contexts in the Ariadne House in the archaeological site of Pompeii, Italy. In particular, (i) the west-facing wall of room 56 located in an outdoor environment was directly exposed for almost two centuries to degradation processes due to weathering and atmospheric pollution that resulted in the total loss of the paint layer; and (ii) the west-faced wall of the basement located under the ground floor below room 56 was constantly sheltered from direct rainwater, sunlight exposure, and drastic thermal fluctuations so that polychromatic decorations are still conserved, although several other degradation evidences were identified. The study was mainly performed using three portable instruments, i.e., a handheld LIBS system incorporating a pulsed Nd:YAG laser operating at 1064 nm and two spectrometers covering the spectral ranges 196–419 nm (UV) and 580–1000 nm (NIR) to

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

489

obtain the elemental characterization of the samples, and two portable Raman systems equipped with a 785- and 532-nm laser, respectively, for the molecular identification of the samples supported by chromatographic studies in the laboratory. The mortars of both walls yielded comparable LIBS data indicating the presence of very intense signals of Ca from CaCO3 used as the binder material, Mg, Al, Si, K, and Na possibly ascribed to volcanic materials and KNO3 soluble salts. The elemental data provided by LIBS were supported by the Raman molecular analyses. The LIBS analysis of pigments in the red and yellow areas on the fresco inside the basement indicated several Fe emission lines in the spectral range between 230 and 390 nm, which suggested iron-based colors. Raman data confirmed the presence of hematite in red areas and goethite in yellow areas. Further, the presence of carbon black mixed with red ochre in the dark red areas and high celadonite content in the green areas were detected by the portable Raman system. The results above proved indirectly that the degradation of the outdoor wall was caused by thermal fluctuations and leaching phenomena caused by the direct exposition to sunlight and acidic rainwater. Further, Raman data showed that the environmental condition of the basement, i.e., high levels of humidity and low exposition to sunlight and temperature fluctuations, determined the formation of several white drips partially covering the mural painting on the basement wall which consisted of CaCO3, of efflorescence salts along the wall surface consisting of NaNO3 and CaSO4 and of green biological patinas on the bottom part of the wall due to the growth of fungal and algal colonies. In conclusion, the results obtained from this unique case of study highlighted the indispensable role of in-situ spectroscopic analysis to understand and predict the degradation pathways jeopardizing the cultural heritage and provide to the Archaeological Park of Pompeii important inference to consider in future conservation projects.

17.3.3 Metal Objects, Coins, Pottery, and Jewelry Multielemental chemical analysis represents an important means for the identification, provenance attribution, type of weathering, and fabrication technologies (archeometallurgy) of archeological metallic artifacts. In particular, the analysis of metal alloys used for the artifact is invaluable for understanding the technology used for its production, dating and origin of the specimen, and the effect of time and environment produced on it. To avoid the use of conventional methods for characterizing artifact materials and possible corrosion layers, which involves sampling from the object and transfer to the laboratory, transportable instruments such as LIBS and XRF represent unique, noninvasive tools that are very appropriate for the in-situ characterization of metal artifacts and may also provide useful information on the sample metal source. In specific situations, however, LIBS can provide more information than XRF, which include LIBS capacity to perform stratigraphic, in-depth compositional analysis able to reveal surface corrosion layers and make remote analysis up to several meters in open path configuration so allowing a much higher spatial resolution. However, the combined use of both portable systems, i.e.,

490

G. S. Senesi and O. De Pascale

LIBS and XRF, should be preferred wherever possible to maximize the information and reduce the impact on the object. Melessanaki et al. [33] used a benchtop LIBS unit based on a compact Q-switched Nd:YAG laser emitting at 1064 nm, which was easily transportable in a van, to identify and quantify a broad variety of archeological findings including glazed pottery, vitreous materials, metal objects, and jewelry. In particular, LIBS analysis of a small bead excavated in a Minoan burial site in Crete revealed it was made of Pb, although its appearance was that of faience. Several pieces of metal jewelry from the late Minoan period (fourteenth century BC) analyzed by LIBS showed emission lines of Au, Ag, and Cu which suggested the use of various proportions of these metals in the alloys. In particular, quantitative LIBS analysis of the amounts of Sn and Cu in a few Minoan bronze objects revealed a content of Sn between 5 and 11%wt, which helped in the classification of the object and determination of the metallurgical capabilities of the site of origin. The LIBS analysis was based on a linear calibration curve constructed by plotting the ratio of emission intensity from Sn and Cu vs the corresponding concentration ratio of several reference bronze samples. In conclusion, LIBS showed to be a powerful tool for performing routine, rapid, on-site sample analysis and/or screening of a large variety of archaeological findings leading to their quick characterization and classification. Bertolini et al. [34] developed the mobile instrument Modì LIBS for in-situ quantitative analysis of various materials including two bronze fragments from a panel from the “Paradise Door” in the Baptistry and an artwork from the church of Orsanmichele, both in Florence, Italy. The innovative experimental setup was based on the use of DP-LIBS pulses and a standardless CF analytical procedure that could overcome problems related to matrix effects and greatly improved the potential of the technique for the accurate quantitative analysis of cultural artifacts. In particular, DP-LIBS typically enhanced the signal on the order of 5- to 50-fold, which improved the LIBS limit of detection of the emitting elements, yielding the quantitative elemental composition in Cu, Sn, Pb, (and Zn) of the two bronze fragments analyzed. Ferretti et al. [35] conducted an in-situ measurement campaign at the National Museum of Magna Grecia, Reggio Calabria, Italy, on several bronze fragments belonging to the so-called “Porticello Bronzes” collection and pertaining to two bronze statues, i.e., the “athlete” and the “philosopher,” dated between the fifth and fourth centuries BC, which were found at sea in an ancient ship sunk in the strait of Messina off the village of Porticello by using two easily hand-portable XRF and Modì DP-LIBS instrumentations. LIBS evidence was obtained of the main components of the alloy, i.e., Cu and Sn; the surface contaminants Mg, K, Na, Ca, Sr, Si, Al, and others possibly originating from the sea environment; and Fe, Ag, Pb, and Bi. In particular, the statistical distribution of the Bi I line at 306.77 nm, Pb I line at 405.78 nm, and Ag I line at 338.29 nm showed significant differences between fragments, which allowed to identify and discriminate all fragments as belonging either to one statue (presence of Bi and relatively intense signals of Ag and Pb) or to the other one (absence of detectable Bi and relatively low signals of Ag and Pb)

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

491

Fig. 17.8 Classification of the fragments in the space of the LIBS signals (Bi, Ag, Pb). The circles correspond to the fragments pertinent to the male nude, while the triangles correspond to the philosopher. (Reprinted from [35], with permission from Elsevier)

(Fig. 17.8). A good agreement was found between LIBS and XRF results that revealed compositional peculiarities that were not observed in past investigations. Agresti et al. [36] developed a novel, compact, low-cost, portable laser-induced plasma spectroscopy (LIPS) system to analyze quantitatively in-situ binary, ternary, and quaternary Cu-alloy artifacts, including a grotesque metal figurine of unknown origin assembled from several pieces from the National Museum of Archeology in Florence, Italy, which is representative of a typical authentication problem. The LIPS system was carefully calibrated using a set of reference samples and applying the internal standard approach, i.e., using the integrated line intensity of the analyte (Sn, Pb, Zn) normalized to the Cu line intensity. Suitable quantification spectral lines of Cu, Sn, Pb, and Zn, which did not exhibit overlapping and whose calibration curves were not affected by matrix effects, were selected preliminarily. Further, a Cl line was selected as an indicator of natural corrosion. Linear calibration curves with regression coefficients above 99% were obtained for Sn, Pb, and Zn contents. The LIPS system reproducibility and repeatability were tested and validated through crossed comparisons with traditional analytical techniques including inductively coupled plasma-optical emission spectroscopy (ICP-OES), atomic absorption spectroscopy (AAS), and SEM-EDX. The elemental depth profile of the Cu-alloy-made

492

G. S. Senesi and O. De Pascale

figurine showed a marked difference between the compositional distribution of the main alloy elements which allowed to clearly discriminate the different degrees of natural aging of the several parts of the artifact. In particular, the arms and feet were crafted using a quaternary alloy typical of the Renaissance art foundry which corroborated a modern integration, whereas the rest of the body featured a ternary composition of the alloy typical of a genuine archeological finding of the Roman period. In conclusion, the instrument and the methodological approach used showed to be a reliable tool to characterize Cu-alloy artifacts and effective to solve a variety of archeometallurgical issues related to composition and corrosion of bronzes and brasses. Osorio et al. [37] used a portable LIBS prototype system emitting at 1064 nm from a multiuse Q-switched Nd:YAG laser for determining the elemental composition of the multilayer structure of a Japanese metal jug from the twentieth century of completely unknown composition belonging to a private collection and highly surface deteriorated. The compositional in-depth profiling of the jar measured at different points showed the presence of two layers, i.e., a surface layer of an estimated 27-μm thickness consisting primarily of Cu and a deeper layer constituted mainly of Pb. The surface layer featured lines of Ca II at 393.36 nm, Si II at 546.68 and 566.95 nm, Na I at 588.99 nm, and K I at 766.48 nm of different and decreasing intensity with depth, which disappeared after four shots and were attributed to dirt and impurity deposition on the sample surface. After 13 pulses all elements, except Pb, virtually disappeared, which indicated the jar was made mainly of this element. Given the presence of a Cu surface coating, the use of acidic solutions (citric acid, acetic acid, sulfuric acid, nitric acid, formic acid) and alkaline solutions (ammonia water, alkaline Rochelle salt, sodium hexametaphosphate, ammonium carbonate) was suggested for chemical restoration. The performance of the portable LIBS prototype used in this study confirmed the known efficiency of LIBS to perform quick, in-situ analysis without the removal of valuable pieces. Grozeva and Penkova [38] reported preliminary qualitative elemental analysis results obtained in-situ by a commercial portable LIBS system on an ancient artifact, i.e., a ritual silver knemida (knee-protecting piece, greave) with many decorations and partial gold plating dated from the fourth century BC and originating from the Mogilanska Tumulus, Vratsa, Bulgaria, a unique example of the Thracian decorative art that belongs to the Thracian museum collection. The system was provided with a handheld head comprising a Q-switched Nd:YAG laser emitting at 1064 nm, which directed the laser beam and collected the plasma light that was transferred to a console containing three spectrometer modules covering the range 250–760 nm. The LIBS elemental composition analysis at characteristic spots on the surface and in depth indicated mainly lines of Cu, Ag, and Au, but also lines of Fe, Mg, Pb, Mn, and Si, whereas Al and Na lines were also identified in the surface layer with intensity decreasing with depth. The low intensity of Cu lines and the presence of Al and Na in the surface layer indicated a previous restoration possibly by air electrochemical treatment. The reported initial measurements were part of a study aiming at the reconstruction of technology of ancient gilding and provided important information for estimating the present condition of the artifact and for choosing the

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

493

appropriate treatment for its restoration and conservation and planning future detailed quantitative analysis of ritual knemida and other objects of the collection. Pardini et al. [39] used a transportable Modì Smart LIBS instrument provided with a DP collinear Nd:YAG laser emitting at the wavelength of 1064 nm, already widely tested for in-situ analysis of art and archeological objects, to analyze in-situ the elemental composition of a large collection of silver Roman Republican denarii encompassing about two centuries of Roman history kept at the “Monetiere” of the Archeological Museum in Florence, Italy. The main elemental lines detected by in-depth LIBS analysis of the coins were those of Ag, Cu, and Pb (Fig. 17.9a). Results obtained by the joint use of portable LIBS and XRF spectroscopy were cross-validated successfully (R ¼ 0.96 for Cu and R ¼ 0.98 for Pb) so confirming the accurate analysis of the coins. Further, the information obtained about the coin composition, both for main components and traces, were used for classifying them in groups according to the different levels of concentrations of the detected elements whose dating was otherwise impossible. In particular, the precise elemental composition of the coins, i.e., the increasing Cu content and decrease of Ag, allowed to reveal a striking correlation between the quality of the silver alloy and the occurrence of some critical contemporary political and military events (Fig. 17.9b). Further, the comparison with other contemporary denarii questioned a recent theory on the reason of issuing the so-called “serrated” denarii (denarii showing notched chisel marks on the edge of the coin). In conclusion, results reported in this work demonstrated the potential of LIBS in analyzing in short time and in-situ a statistically significant number of coins for dating and authentication issues and suggested the possibility of extending the methodology used in this study to other types of coins. Gaona et al. [23] analyzed by LIBS the elemental composition of two decorative metallic details, i.e., the flowers in the vases alongside the medallion and the seals in the corners of the squared frame, located on the façade of the Cathedral of Malaga, Spain. LIBS spectra of surface polluted flowers after the first shot showed a strong emission line of Pb at 405.8 nm and the presence of lines from Mg, Fe, Si, Ca, Al, Ti, Mn, and Sr, which disappeared completely after 300 shots on the same spot. These

Fig. 17.9 (a) LIBS spectrum of a Republican denarius. (b) Time evolution of the copper content in silver denarii. (Reprinted from [39], with permission from Elsevier)

494

G. S. Senesi and O. De Pascale

results indicated that flowers were made exclusively by Pb, whereas all other pollutant elements covering the surface originated from natural sources, i.e., wind blowing from the nearby coast, and anthropogenic sources, mainly road traffic in urban area. Differently, the seals featured atomic and ionic Cu emissions with small contribution of Pb, which suggested they were made of a leaded copper bronze alloy. Siano and Agresti [40] investigated by a built-in-house portable LIPS device equipped with a Q-switched Nd:YAG laser emitting at 1064 nm the composition and the execution processes of a set of well-known Florentine Cu-alloy masterpieces attributed to Donatello, i.e., the David, San Ludovico, and Pulpito della Resurrezione, executed through multiple casting and assembling. The quantitative elemental in-depth profiles acquired from several parts of the three masterpieces allowed to measure in detail the amounts of the major elements, i.e., Cu, Pb, Sn, and Zn, and the predominant traces constituting the various bronze and brass alloys of which the artifacts were made and highlight the casting technology used. The detection limits of Sn, Zn, and Pb were about 0.1 wt%. In particular, the San Ludovico masterpiece, known as the first Donatello Cu-alloy sculpture, was found to be cast in several pieces, mostly using a quaternary alloy (Cu, Zn, Sn, Pb), i.e., a leaded low-Zn brass, which were separately gilded and then assembled mechanically. The LIPS analysis of the David, Donatello most celebrated masterpiece, showed that the two main pieces cast together, i.e., the base and the two figures, and all sculptural elements and repairs were made by ternary Cu alloys containing 3.4% Sn and 4.1% Pb. The LIPS analysis of decorative panels with reliefs of the parapets of the two pulpits forming the “Pulpito della Resurrezione” in the Basilica of S. Lorenzo in Florence showed that they were assembled from several pieces all cast using leaded bronze with an almost constant content of Sn and presence of relatively high contents of Fe (2–5%) and Ni (0.5–2%). In conclusion, two different types of alloy were revealed by LIPS, i.e., a low-Zn brass in San Ludovico and Pb bronzes in the David and Pulpito della Resurrezione. LIPS data integrated with ICP-OES and AAS data achieved on other Cu-alloy masterpieces of the Early Renaissance Florentine production allowed to draw a comprehensive compositional and metallurgical picture of Cu-alloy recipes in use in art foundries during that time to execute very complex sculptures by using several independent casting operations. Recently, Brysbaert et al. [41] used a compact portable LIBS instrument based on a passively Q-switched Nd:YAG laser operating at 1064 nm equipped with a spectrometer covering the range from 200 to 660 nm for the in-situ elemental depth-profiling stratigraphic analysis of about 150 different pyrotechnological materials, including ceramic crucibles, Cu-alloy and Pb objects, and glass beads and faience fragments from several late Bronze Age (mid- and late thirteenth century BC) at workshop areas located at the archeological site of Tiryns, Greece. Strong emissions from Cu, alone or together with weak lines of Sn, were detected in most crucible deposits, which suggested their main use for the preparation of pure Cu or bronze items. In a few cases, a weak signal of Pb (and Ag) was detected suggesting the addition of Pb to the Cu alloy to improve castability and reduce the melting temperature. Further, the clear signal of Fe found in many crucibles might indicate a clay contamination. Either a strong signal of Cu alone or a Cu-Sn-alloy spectral line

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

495

pattern was detected in over 60 Cu-alloy objects and melting wastes, such as scarp and spills, which are typical of prehistoric alloys. Some samples also showed the presence of Pb (and Ag), which suggested bronze alloys, and presence of Fe possibly due to surface contamination. The presence of V in several of the 30 lead objects analyzed was ascribed tentatively to clay or volcanic ash contamination. A variety of metallic elements implying the presence of multiple colorants were detected in 22 glass and faience beads and rhyton fragments. In particular, Mn and Cu were detected at various intensities in most of them, indicating the presence of manganese black or purple/violet hues and copper green-blue, respectively. The detection of Pb in some items suggested the use of lead antimonite to produce a yellow glaze line or, in combination with Cu, a green glaze line. Finally, the presence of Co in two glass beads indicated its use to produce a blue-black line decoration. In conclusion, the findings of this work allowed to identify unknown materials, suggest their provenance, and trace their production processes.

17.3.4 Submerged Materials Submerged materials represent a very valuable source of information on the origin of the wrecks and materials employed in manufacturing the artifacts found in them. However, very often the study of archeological materials from marine environment is not practical due to the size of the sample, or their removal is not permitted by legislation or preservation practices. Thus, the analysis in the same place where the archeological materials are discovered, i.e., in-situ, represents the only alternative to measure the chemical composition and obtain information of the sample of interest. Actually, no other analytical technique, except LIBS, appears able to provide the atomic composition of a sample submerged at depths of tens of meters in a real sea environment. Guirado et al. [42] developed and adapted to the marine environment a mobile remote LIBS instrument for the recognition and identification of archeological materials, including ceramic pottery, bones, and metallic samples submerged in seawater both in laboratory-simulated conditions and in on-site trials at depths down to 30 m. The submersible LIBS instrument was based on a 45-m-long fiber optic cable able to transmit the laser beam energy to the handheld probe operated by a diver. An air flux was supplied by a compressor to the probe to remove the water from the sample surface, create a gas-solid sample interface prior to laser ablation, and prevent the entrance of water into the LIBS probe. This flux prevented the contact of seawater with the sample surface during the analysis, thus improving the ablation efficiency and allowing the acquisition of good-quality LIBS spectra. The plasma light was collected through the same optical fiber and returned to the data acquisition module placed on the research vessel. To optimize the best conditions for field analysis, the gas flow pressure, beam focal conditions, and angle of incidence were checked preliminarily in simulated laboratory conditions. In order to simulate the real marine environment conditions, LIBS spectra in the 350–550-nm range were acquired on a number of materials submersed inside a tank filled with seawater in the

496

G. S. Senesi and O. De Pascale

laboratory. Results showed that an archeological pottery was composed mainly of quartz (SiO2), clay (Al silicate of Ca and Mg), TiO2, and Fe oxides, whereas a bone sample presented intense emissions from Ca and Sr and trace levels of Fe. Further, Ag and Au lines were detected clearly in precious metal samples and concretion layers deposited on sample surfaces composed mainly of Fe. Some samples, i.e., various bronzes with a high corrosion and oxidation degree, were analyzed directly underwater at 30-m depth during a campaign in the Malaga Bay in the Mediterranean Sea. Despite the poor visibility, good-quality LIBS spectra could be acquired which showed the main emission lines of Cu I at 521.93 nm, Pb I at 405.89, nm and Zn (Fig. 17.10). The identification of a bronze with a higher content of Zn was discriminated from the other two pieces where its content was almost negligible. One piece of these showing a stronger Pb line was identified as a leaded bronze. The dependence of LIBS signal on the immersion depth was also studied on the leaded bronze that showed a progressive decrease of the Pb signal with increasing depth, whereas the Cu signal remained about constant. In conclusion, the mobile remote LIBS device showed very promising for the in-situ characterization of objects in underwater archaeological sites, although the immersion depth would affect the signal from some elements, determine their geographical origin, and understand the historical context.

Fig. 17.10 LIBS spectra of bronze alloys acquired at a 30-m depth: (a) bronze with a higher content of Zn, (b) archeological bronze, and (c) bronze with a leaded matrix. (Reprinted from [42], with permission from Elsevier)

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

497

In a successive paper, Guirado et al. [43] introduced a novel, more compact, and robust remote LIBS prototype named “Aqualas 2.0” that used both SP and multipulse (MP) excitation for chemical analysis of submerged materials. The system consisted of two parts, i.e., a main unit, comprising the optical module and the data acquisition module and the laser power supply which was deployed on the vessel board and was interconnected by means of a 50-m-long umbilical cable to a handheld probe operated by a diver. The new LIBS system represented a technological evolution of the one used previously [42] in that it allowed the transmission of a higher laser radiation through the optical fiber, thus improving the performance of the equipment in terms of range of analysis (down to a depth of 50 m) and variety of samples to be analyzed. Laboratory experiments performed on an ancient pottery covered by both calcareous and ferrous deposits immerged in a water tank simulating real conditions yielded the spectral fingerprints of both deposits, i.e., lines of Ca, Mg, and Sr and traces of Fe for the first one and of Fe with Ca and Mg as minor elements for the other deposit. Further, a measurement campaign was performed in a real underwater archeological site in the Bay of Cadiz, Atlantic Ocean, on a cannon and other well-preserved objects including pieces of copper, iron, and lead, found in the wreck of the “Bucentaure” ship at 17-m depth, after removing the concretion layer from the sample surfaces. The LIBS analysis of the cannon confirmed the relevant presence of Fe, whereas other objects, such as a rosary and a small metallic piece, were identified as iron alloys. In conclusion, the results obtained in this work confirmed the maturity of LIBS to perform in a marine environment. Further, the technological evolution reached by LIBS analysis in the MP collinear configuration, besides reducing the dimensions, consumption, and equipment costs, improved markedly the performance of the remote LIBS instrument in terms of energy transmitted through the optical fiber; intensity enhancement by a factor of 15, if compared to single pulse (SP)-LIBS; range of analysis; and variety of materials to be analyzed, including marble and ceramics, besides metals and alloys. Recently, a new-generation remote LIBS instrument was used by Lopez-Claros et al. [44] to acquire LIBS spectra providing a chemical fingerprint able to identify ancient artifacts from shipwrecks. The instrument consisted of a main unit comprising the optical module containing a Q-switched Nd:YAG laser operating at 1064 nm, which launched the laser beam into a 55-m-long optical cable that transmitted it to the interconnected handheld probe that focused it onto the sample surface and collected and returned the plasma light emitted from the sample to the data acquisition module. Preliminarily, a linear discriminant analysis (LDA) model was constructed based on the spectra acquired in the laboratory in conditions simulating a submerged sea environment from a set of 38 known archeological objects collected from several shipwrecks. The objects subjected to chemometric analysis were divided into four distinct groups, i.e., bronze alloys, metallic pieces, ceramics, and marbles, the LIBS spectra were acquired in the 350–550-nm spectral range, and the peak intensities were normalized to unity for comparison. By using only ten spectral emission lines, i.e., those of Cu I at 510.55 nm, Zn I at 481.05 nm, Sn I at 452.47 nm, Pb I at 405.78 nm, Fe I at 438.35 nm, Ca I at 422.67 nm, Mg I at 517.26 nm, Si I at

498

G. S. Senesi and O. De Pascale

390.55 nm, Sr I at 407.61 nm, and Ti I at 498.17 nm, all samples could be correctly identified and separated into the expected four distinct classes. Further, the model tested in subsea-simulated laboratory conditions on a set of sheathing pieces collected from different shipwrecks from four underwater archeological sites of different age was able to sort and classify the samples chronologically and geographically on the basis of the evolution of their elemental composition from lead to copperbased alloys (Fig. 17.11). Finally, a field campaign was performed on the wreck of San Pedro de Alcantara located in the coastal area of South of Spain Mediterranean Sea at a depth of 10 m over a sandy bottom surface. The main unit of the LIBS instrument was deployed on the vessel board, while a diver operated the handheld probe in the sea bottom using an Ar purge gas that provided LIBS signals more intense than air. The LIBS-LDA analysis of the large number of archeological objects discovered after removing sediments, calcareous deposits, and marine algae covering the structure allowed to classify them unequivocally into four groups, i.e., bronze alloys, ceramic fragments, metallic pieces, and marbles. Further, it was possible to locate the archeological pieces on the wreck (Fig. 17.12). Thus, remote LIBS devices were confirmed to be a unique and powerful tool for inspecting materials in underwater archeology.

Fig. 17.11 Chronocultural sorting of evolution of sheathing’s composition from lead to copperbased alloys. (Reprinted from [44], with permission from Elsevier)

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

499

Fig. 17.12 Schematic drawing of the shipwreck. The inset bar diagrams represent the chemical composition of each object and the locations of the archaeological findings. (Reprinted from [44], with permission from Elsevier)

500

17.4

G. S. Senesi and O. De Pascale

Conclusions and Perspectives

In the recent years, the performance of mobile instrumentations has grown and improved markedly as a tool for in-situ cultural heritage investigations allowing contactless, trustable, fast, sensitive, multielemental, noninvasive, non- or microdestructive, minimum impact analysis of art objects. LIBS applications in cultural heritage and archaeology, if applied correctly, have shown able to yield unique information instrumental to provide answers to relevant problems in the field, such as authentication, dating, and conservation. In particular, results reviewed in this chapter recently obtained in-situ by the use only of mobile, i.e., transportable, portable, and handheld, LIBS instruments on various ancient artifacts appear very promising. Further, the LIBS technique shows unique promising features, not only for the in-situ analysis but also for remote, simultaneous, and multielemental analysis at distances of up to tens of meters. This aspect is very important, in particular in investigations of historical buildings and monuments and in archeological excavation sites. In particular, handheld LIBS instrumentation recently introduced in the cultural heritage field appears to be very promising for performing in-situ measurements on outdoor monuments, archaeological sites, and mural and cave paintings, being able to classify objects, identify surface alterations, and decide previously any further intervention at the conservation laboratory. However, more work is needed to increase the performance of handheld instruments and lower their cost, thus making more attractive the routinary use of LIBS for direct on-site measurements, also in order to obtain quantitative analytical data. In particular, the availability of proper custom-made software able to provide spectral line data for each element and enable simulation of emission spectra and/or the use of reference spectra for immediate comparison with experimental LIBS spectra would allow a relatively simple and rapid identification of the elemental composition of the samples studied also by a nonspecialized user. In the future, the use of appropriate data correlation methods and chemometrics is expected to enable appropriate multivariate statistical LIBS data processing that can provide the classification and provenance of the raw parent materials used to build the artifact of interest on the basis of spectral similarities and differences. Further, the knowledge of the detailed elemental composition of art objects would provide archaeologists, historians, and conservators unique information on the historical context and chronological period to which the object studied belongs. In conclusion, the future of LIBS in cultural heritage and archaeology studies on site relies mainly on the development of cheaper and more performing mobile LIBS systems, possibly coupled to other complementary mobile instrumentations such as XRD and Raman spectroscopy, and on the full comprehension and consequent exploitation of easy solutions to enhance the emission signal of the spectra and reduce further the destructivity of the technique.

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

501

References 1. Vandenabeele P, Donais MK (2016) Mobile spectroscopic instrumentation in Archaeometry research. Appl Spectrosc 70(1):27–41 2. Giakoumaki A, Melessanaki K, Anglos D (2007) Laser-induced breakdown spectroscopy (LIBS) in archaeological science-applications and prospects. Anal Bioanal Chem 387:749–760 3. Anglos D, Detalle V (2014) Cultural heritage applications of LIBS. In: Musazzi S, Perini U (eds) Laser-induced breakdown spectroscopy. Springer, Berlin, pp 531–554 4. Spizzichino V, Fantoni R (2014) Laser induced breakdown spectroscopy in archeometry: a review of its application and future perspectives. Spectrochim Acta Part B 99:201–209 5. Lazic V, Vadrucci M, Fantoni R, Chiari M, Mazzinghi A, Gorghinian A (2018) Applications of laser-induced breakdown spectroscopy for cultural heritage: a comparison with X-ray fluorescence and particle induced X-ray emission techniques. Spectrochim Acta Part B 149:1–14 6. Botto A, Campanella B, Legnaioli S, Lezzerini M, Lorenzetti G, Pagnotta S, Poggialini F, Palleschi V (2019) Applications of laser-induced breakdown spectroscopy in cultural heritage and archaeology: a critical review. J Anal At Spectrom 34:81–103 7. Gobernado-Mitre I, Prieto AC, Zafiropulos V, Spetsidou Y, Fotakis C (1997) On-line monitoring of laser cleaning of limestone by laser-induced breakdown Spectroscopy and laser-induced fluorescence. Appl Spectrosc 51:1125–1129 8. Maravelaki-Kalaitzakia P, Anglos D, Kilikoglou V, Zafiropulos V (2001) Compositional characterization of encrustation on marble with laser induced breakdown spectroscopy. Spectrochim Acta Part B 56:887–903 9. Ortiz P, Vázquez MA, Ortiz R, Martin JM, Ctvrtnickova T, Mateo MP, Nicolas G (2010) Investigation of environmental pollution effects on stone monuments in the case of Santa Maria La Blanca, Seville (Spain). Appl Phys A Mater Sci Process 100:965–973 10. Senesi GS, Carrara I, Nicolodelli G, Milori DMBP, De Pascale O (2016) Laser cleaning and laser-induced breakdown spectroscopy applied in removing and characterizing black crusts from limestones of Castello Svevo, Bari, Italy: a case study. Microchem J 124:296–305 11. Senesi GS, Nicolodelli G, Milori DMBP, De Pascale O (2017) Depth profile investigations of surface modifications of limestone artifacts by laser-induced breakdown spectroscopy. Environ Earth Sci 76:565 12. Cremers DA, Radziemski LJ (eds) (2006) Handbook of laser-induced breakdown spectroscopy, 2nd edn. Wiley, Chichester 13. Senesi GS, Harmon RS, Hark RR (2021) Field-portable and handheld laser-induced breakdown spectroscopy: historical review, current status and future prospects. Spectrochim Acta Part B 175:106013 14. Noll R (ed) (2012) Laser-induced breakdown spectroscopy fundamentals and applications, 2nd edn. Springer, Berlin/Heidelberg 15. Cristoforetti G, De Giacomo A, Dell’Aglio M, Legnaioli S, Tognoni E, Palleschi V, Omenetto N (2010) Local thermodynamic equilibrium in laser-induced breakdown spectroscopy: beyond the McWhirter criterion. Spectrochim Acta Part B 65:86–95 16. Fortes FJ, Laserna JJ (2010) The development of fieldable laser-induced breakdown spectrometer: no limits on the horizon. Spectrochim Acta Part B 65:975–990 17. Rakovský J, Cermak P, Musset O, Veis P (2014) A review of the development of portable laser induced breakdown spectroscopy and its applications. Spectrochim Acta Part B 101:269–287 18. Dequaire T, Meslin P-Y, Beck P, Jaber M, Cousin A, Rapin W, Lasne J, Gasnault O, Maurice S, Buch A, Szopa C, Coll P (2017) The MSL Science Team: analysis of carbon and nitrogen signatures with laser-induced breakdown spectroscopy; the quest for organics under Mars-like conditions. Spectrochim Acta Part B 131:8–17 19. Ciucci A, Corsi M, Palleschi V, Rastelli S, Salvetti A, Tognoni E (1999) New procedure for quantitative elemental analysis by laser induced plasma spectroscopy. Appl Spectrosc 53: 960–964

502

G. S. Senesi and O. De Pascale

20. Grönlund R, Lundqvist M, Svanberg S (2005) Remote imaging laser-induced breakdown spectroscopy and remote cultural heritage ablative cleaning. Opt Lett 30:2882–2884 21. Fortes FJ, Cunat J, Cabalin LM, Laserna JJ (2007) In situ analytical assessment and chemical imaging of historical buildings using a man-portable laser system. Appl Spectrosc 61:558–564 22. Brai M, Gennaro G, Schillaci T, Tranchina L (2009) Double pulse laser induced breakdown spectroscopy applied to natural and artificial materials from cultural heritages: a comparison with micro-X-ray fluorescence analysis. Spectrochim Acta Part B 64:1119–1127 23. Gaona I, Lucena P, Moros J, Fortes FJ, Guirado S, Serrano J, Laserna JJ (2013) Evaluating the use of standoff LIBS in architectural heritage: surveying the Cathedral of Malaga. J Anal At Spectrom 28:810–820 24. Senesi GS, Campanella B, Grifoni E, Legnaioli S, Lorenzetti G, Pagnotta S, Poggialini F, Palleschi V, De Pascale O (2018a) Elemental and mineralogical imaging of a weathered limestone rock by double-pulse micro-laser-induced breakdown spectroscopy. Spectrochim Acta Part B 143:91–97 25. Senesi GS, Manzini D, De Pascale O (2018b) Application of a laser-induced breakdown spectroscopy handheld instrument to the diagnostic analysis of stone monuments. Appl Geochem 96:87–91 26. Aramendia J, Gómez-Nubla L, Fdez-Ortiz de Vallejuelo S, Castro K, Arana G, Madariaga JM (2019) The combination of Raman imaging and LIBS for quantification of original and degradation materials in cultural heritage. J Raman Spectrosc 50:193–201 27. Westlake P, Siozos P, Philippidis A, Apostolaki C, Derham B, Terlixi A, Perdikatsis V, Jones R, Anglos D (2012) Studying pigments on painted plaster in Minoan, Roman and Early Byzantine Crete. A multi-analytical technique approach. Anal Bioanal Chem 402:1413–1432 28. Kaszewska EA, Sylwestrzak M, Marczak J, Skrzeczanowski W, Iwanicka M, Szmit-Naud E, Anglos D, Targowski P (2013) Depth-resolved multilayer pigment identification in paintings: combined use of laser-induced breakdown spectroscopy (LIBS) and optical coherence tomography (OCT). Appl Spectrosc 67:960–972 29. Papliaka ZE, Philippidis A, Siozos P, Vakondiou M, Melessanaki K, Anglos D (2016) A multitechnique approach, based on mobile/portable laser instruments, for the in situ pigment characterization of stone sculptures on the island of Crete dating from Venetian and Ottoman period. Herit Sci 4(15):1–18 30. Siozos P, Philippidis A, Anglos D (2017) Portable laser-induced breakdown spectroscopy/ diffuse reflectance hybrid spectrometer for analysis of inorganic pigments. Spectrochim Acta Part B 137:93–100 31. Lazic V, Vadruccia M, Fantoni R, Chiari M, Mazzinghi A, Gorghinian A (2018) Applications of laser-induced breakdown spectroscopy for cultural heritage: a comparison with X-ray fluorescence and particle induced X-ray emission techniques. Spectrochim Acta Part B 149:1–14 32. Veneranda M, Prieto-Taboada N, Fdez-Ortiz de Vallejuelo S, Maguregui M, Morillas H, Marcaida I, Castro K, Garcia-Diego F-J, Osanna M, Madariaga JM (2018) In-situ multianalytical approach to analyze and compare the degradation pathways jeopardizing two murals exposed to different environments (Ariadne House, Pompeii, Italy). Spectrochim Acta Part A 203:201–209 33. Melessanaki K, Mastrogiannidou A, Chlouveraki S, Ferrence SC, Betancourt PP, Anglos D (2005) Analysis of archaeological objects with LMNTI, a new transportable LIBS instrument. Springer Proc Phys 100:443–449 34. Bertolini G, Carelli F, Francesconi M, Francesconi L, Marchesini P, Marsili F, Sorrentino G, Cristoforetti S, Legnaioli V, Palleschi L, Pardini A, Salvetti A (2006) Modì: a new mobile instrument for in situ double-pulse LIBS analysis. Anal Bioanal Chem 385:240–247 35. Ferretti M, Cristoforetti G, Legnaioli S, Palleschi V, Salvetti A, Tognoni E, Console E, Palaia P (2007) In situ study of the Porticello Bronzes by portable X-ray fluorescence and laser-induced breakdown spectroscopy. Spectrochim Acta Part B 62:1512–1518 36. Agresti J, Mencaglia AA, Siano S (2009) Development and application of a portable LIPS system for characterising copper alloy artefacts. Anal Bioanal Chem 395:2255–2262

17

Laser-Induced Breakdown Spectroscopy (LIBS) In-Situ: From Portable to. . .

503

37. Osorio LM, Cabrera LVP, Garcia MAA, Reyes TF, Ravelo I (2011) Portable LIBS system for determining the composition of multilayer structures on objects of cultural value. J Phys Conf Ser 274(1–6). 012093/012091–012093/012096 38. Grozeva M, Penkova P (2012) In situ LIBS analysis of valuable museum objects: the Mogilanska tumulus ritual Knemida. In: Rădvan R, Akyüz S, Simileanu M (eds) The unknown face of the artwork. İstanbul Kültür University, İstanbul, pp 103–107 39. Pardini L, El Hassan A, Ferretti M, Foresta A, Legnaioli S, Lorenzetti G, Nebbia E, Catalli F, Harith MA, Diaz Pace D, Anabitarte Garcia F, Scuotto M, Palleschi V (2012) X-ray fluorescence and laser-induced breakdown spectroscopy analysis of Roman Silver Denarii. Spectrochim Acta Part B 74–75:156–161 40. Siano S, Agresti J (2015) Archaeometallurgical characterisation of Donatello’s Florentine copper alloy masterpieces using portable laser-induced plasma spectroscopy and traditional techniques. Stud Conserv 60:S106–S119. https://doi.org/10.1179/0039363015Z. 000000000215 41. Brysbaert A, Siozos P, Vetters M, Philippidis A, Anglos D (2017) Materials analyses of pyrotechnological objects from LBA Tiryns, Greece, by means of laser-induced breakdown Spectroscopy (LIBS): results and a critical assessment of the method. J Archaeol Sci 83:49–61 42. Guirado S, Fortes FJ, Lazic V, Laserna JJ (2012) Chemical analysis of archaeological materials in submarine environments using laser-induced breakdown spectroscopy. On-site trials in the Mediterranean Sea. Spectrochim Acta Part B 74–75:137–143 43. Guirado S, Fortes FJ, Laserna JJ (2015) Elemental analysis of materials in an underwater archeological shipwreck using a novel remote laser-induced breakdown spectroscopy system. Talanta 137:182–188 44. López-Claros M, Fortes FJ, Laserna JJ (2018) Subsea spectral identification of shipwreck objects using laser-induced breakdown spectroscopy and linear discriminant analysis. J Cult Herit 29:75–81

Provenance of Italian and Central European Archaeological Obsidians by Non-destructive WDXRF Method

18

A. M. De Francesco, M. Bocci, and G. M. Crisci

Contents 18.1 18.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Distribution of the Obsidian Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.1 Central-Eastern Mediterranean Obsidian Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.2.2 Carpathian Obsidian Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.3 The Analyzed Archaeological Obsidian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.4 Nondestructive XRF Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18.5 Results and Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

506 507 507 508 509 509 511 515

Abstract

This chapter deals with the provenance of 1453 archaeological obsidian fragments from Italian and Central European sites, determined by the nondestructive wavelength-dispersive X-ray fluorescence (WDXRF) analytical method, proposed by Crisci et al. (Gallia Préistoire 36:299–327, 1994) and optimized by De Francesco et al. (Archaeometry 50(2):337–350, 2008). In the initial phase, this methodology was applied to Mediterranean obsidians where numerous geological obsidian sources are present. Indeed representative obsidian samples were collected from all the geological outcrops and analyzed for major and trace elements by traditional XRF method on powders. Successively, the entire obsidian fragments (with shape similar to archaeological waste) removed from the same geological samples were analyzed by nondestructive XRF method, by using the secondary X-ray intensities of selected trace elements such as Nb, Y, Zr, Rb, and Sr, instead of their absolute concentrations, as described in De Francesco et al. (Archaeometry 50(2):337–350, 2008).

A. M. De Francesco (*) · M. Bocci · G. M. Crisci Dipartimento di Biologia, Ecologia e Scienze della Terra, Università della Calabria, Rende, CS, Italy e-mail: [email protected]; [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_18

505

506

A. M. De Francesco et al.

The comparison between the results of the two methodologies shows that the X-ray intensity ratios are able to effectively discriminate between the obsidian sources of the whole Mediterranean area, in the same way as with concentrations on powders. The selected five chemical elements, due to their significant role in genetic processes, are more than enough to characterize the different obsidian sources in the Mediterranean area. The provenance of the unknown obsidian artifacts was determined by comparing their trace elements X-ray intensity ratios with those obtained on the entire fragments of the obsidian sources. With the nondestructive WDXRF method, in about two decades, thousands of archeological obsidian have been analyzed and assigned to their own obsidian source. In the present work, the results obtained on 1453 archaeological obsidians from numerous Italian, Corsican, Central European (Romanian) archaeological sites are shown. The provenance of the 97% of the obsidian archaeological fragments was successfully determined. The nondestructive XRF method used has proved extremely effective for attributing the origin of archaeological obsidians in the considered areas.

18.1

Introduction

Studies on the provenance of archaeological obsidians, through the geochemical fingerprinting of the obsidian sources, have been carried out since the 1960s of the last century, by using different methods (destructive and nondestructive): optical emission spectroscopy [13]; instrumental neutron activation analysis [28, 30, 47]; WDXRF on powders by Francaviglia [26]; energy-dispersive X-ray spectrometry coupled with scanning electron microscopy [1, 2, 11]; fission track dating [8]; Williams-Thorpe [46] by different methodologies; Kayani and McDonnel [31] by backscattered electron-scanning electron microscopy; inductively coupled plasma mass spectrometry [44] and Tykot [42]; gamma rays XRF and other different methodologies [36–39]; proton-induced X-rays and gamma rays [41]; inductively coupled plasma mass spectrometry coupled with laser ablation [3, 24, 29]; Mossbauer spectroscopy [40]; Raman microspectroscopy, PIXE, and ICP [5, 6]; and portable X-ray fluorescence spectrometer [43]. The different sources of obsidian, due to the imprint of genetic processes, are characterized by a well-defined chemical composition of the trace elements. Nevertheless, archaeological obsidians cannot be destroyed; so to analyze them the problem has always been a methodological one. In the studies on the provenance of archaeological obsidians, it is very important to analyze all the fragments found in an excavation, because even only one of them with a different origin can make a great contribution to understanding the trade routes of the investigated populations. For these reasons it is essential to analyze all

18

Provenance of Italian and Central European Archaeological Obsidians by. . .

507

the archaeological obsidian fragments, without any preparation or damage and also with low costs. In order to preserve the integrity of the archaeological obsidians, a WDXRF nondestructive analytical methodology has been introduced by Crisci et al. [14] and optimized by De Francesco et al. [21]. For over 20 years, with the nondestructive WDXRF method, numerous geological and archaeological obsidians have been analyzed from the Mediterranean area [15, 19, 22, 23], central-western Argentina and central Chile in South America [20, 24, 25], and Carpathian Central European obsidian sources [7]. In this chapter, about 1453 obsidian fragments from numerous archaeological sites in Italy, Corsica, and Central Europe (Romania) were analyzed and ascribed to their geological sources.

18.2

Distribution of the Obsidian Sources

Obsidian-bearing volcanic are found only in restricted areas of the world. In particular, the European continent is an excellent region for obsidian studies because only in some Italian and Aegean islands and in northeastern Hungary (Tokaj Mountains) and eastern Slovakia (Zemplin Hills) in the Carpathian area [33] obsidian suitable for toolmaking is present.

18.2.1 Central-Eastern Mediterranean Obsidian Sources The geological obsidian sources of the Mediterranean area, exploited in the Neolithic age, are distributed in the Italian islands of Sardinia, Lipari, Palmarola, and Pantelleria and in the Greek islands of Melos and Gyali (Fig. 18.1). Here, the obsidian outcrops are very numerous, often very small sized and not always of glassy and highly translucent quality. In the Monte Arci region of Sardinia, Hallam et al. [30] and Machey and Warren [32] distinguished three obsidian sources: Conca Cannas (SA), Santa Maria Zuarbara (SB), and Perdas Urias (SC). In the following years, Williams-Thorpe et al. [47], Tykot [42], and De Francesco et al. [15, 16, 18, 21], through several analytical methods, have identified up to five geochemically separated groups (SA, SB1, SB2, SC1, and SC2). On Lipari, one of the Aeolian islands, Gabellotto obsidian flow is the only one exploited in the Neolithic period. The Palmarola island is part of the Pontine archipelago, located in the Tyrrhenian Sea, off the coast of the Gulf of Gaeta (southern Lazio region). On this island two obsidian kinds occur, one transparent with inclusions and one black and opaque. Tykot et al. [45] recognized three sub-sources on Palmarola island, through the NAA method.

508

A. M. De Francesco et al.

Fig. 18.1 Distribution of the obsidian sources in the Mediterranean area and in the Carpathian Mountains (Central Europe)

Pantelleria is a Sicilian island located in the center of the Sicilian Channel. In Pantelleria three well-separated geochemical groups of obsidian have been distinguished: Balata dei Turchi, Salto La Vecchia, and Lago di Venere [27]. Due to the peralkaline composition, these obsidians have both macroscopic (greenish color) and geochemical features that make them particularly distinguishable in the Mediterranean area. In the Aegean Sea, the main sources of supply were constituted by Giali, on the Anatolian coast, and by Milos, in the Cyclades. The Milos and Giali obsidians are both of good quality and chemically distinct [13, 26, 34]. Also on Antiparos, another island in the Greek archipelago, there is obsidian but of poor quality, and therefore it has never been used as a raw material.

18.2.2 Carpathian Obsidian Sources In the Central Europe, the obsidian sources occur in the Carpathian Basin (Fig. 18.1), on the Hungarian-Slovakian border [33]. The Carpathian obsidian, although produced by multiple eruptions, is in a small quantity; however, these obsidians are very important because they are the unique natural sources of raw material recognized in Central Europe. In fact the sources found near Vinicky and Cejkov in southeastern Slovakia gave good-quality obsidian and so were exploited in Neolithic period [9, 10, 12].

18

Provenance of Italian and Central European Archaeological Obsidians by. . .

509

Carpathian obsidians are found in Hungary and Ukraine. In Hungary, the main obsidian outcrops are in the southern and central ranges of the Tokaj Mountains (Mad and Tolcsva environs), while in Ukraine it is present in Transcarpathia [33, 35]. Carpathian obsidians could be divided into two major geochemical groups termed Carpathian 1 (C1) found near the towns of Vinicky and Cejkov (Slovakia) according to Williams-Thorpe et al. [47] and Carpathian 2 (C2) from the Tokaj Mountains, near Tolcsva and Erdobénye (Hungary). The compositional data by NAA [33] confirmed the two chemical groups identified by Williams-Thorpe et al. [47], C2A and C2B subgroups. The NAA analysis was able to distinguish two subgroups of Cl based on concentrations of Rb, U, Sb, and Se.

18.3

The Analyzed Archaeological Obsidian

In this chapter 1453 archaeological obsidian fragments from several Italian, Corsican, and Central European archaeological sites were analyzed using the nondestructive XRF. For most archaeological sites, all the obsidian fragments found in the excavation were analyzed; in other cases, archaeologists have selected representative samples, as very experienced eyes can detect similarities and differences or recognize the origin of some obsidians only by macroscopic examination. Archaeological obsidians from about 1 cm to 5 cm can be analyzed. This limit is determined by the size of the sample container diameter of the employed XRF spectrometer. The X-rays irradiated area varies from 1 to 2 cm in diameter and is proportional to the fragment size. The geographical location of the obsidian fragments analyzed in this chapter is divided into the following areas: 1398 obsidian fragments from 45 archaeological sites in Central, Southern, and Northern Italy and Sardinia-Corsica and 55 obsidian fragments from 7 archaeological sites in Central Europe-Carpathian Mountains. More detailed information on the archaeological sites investigated can be found in De Francesco et al. [23] and Biagi et al. [7] for Mediterranean and Carpathian areas, respectively.

18.4

Nondestructive XRF Methodology

All the obsidian sources of the Mediterranean [21] were analyzed by both traditional XRF and nondestructive XRF methods by using Philips PW 1480 spectrometer. Carpathian obsidians were analyzed only by nondestructive XRF [7]. In the preliminary phase of the work, the geological obsidian sources were sampled and analyzed by XRF analysis on the powdered samples by Philips PW 1480 spectrometer. This analysis allowed us an unequivocal characterization of the different obsidian sources. The nondestructive XRF method, optimized at the Dipartimento di Scienze della Terra of the University of Calabria [14, 21], allows to trace the origin of

510

A. M. De Francesco et al.

the archaeological obsidians by directly analyzing the entire artifacts, without any sample preparation and leaving the analyzed sample perfectly intact. The nondestructive XRF method, described in detail in De Francesco et al. [21] and summarized here, consists of irradiating with primary X-rays entire splinters of obsidian taken from the larger blocks, collected from the geological outcrops (already analyzed as powders), similar in shape to workshop refuse (Fig. 18.2), usually found at archaeological sites, using the same Philips PW 1480 spectrometer. As a result, the chemical elements present in the sample are excited and produce secondary X-rays. In the traditional XRF, the matrix effects (as the absorption, the granulometric, and the surface effects) must be considered; but analyzing the archaeological obsidians with the nondestructive XRF method, the only theoretical obstacle is linked to the surface irregularity which is also different for each fragment. If we consider that the radiation emitted from two similar and simultaneously analyzed chemical elements in the same sample suffers similar variations, then the secondary X-ray intensity ratios of selected chemical elements can be used in place of real concentrations and, therefore, employed to construct discriminating diagrams [21]. Among all the elements analyzed, only Nb, Y, Zr, Rb, and Sr were selected, due to their relevance in the magmatic processes that lead to the fusion of rhyolitic magmas which solidifying, give rise to obsidian, and are more than sufficient to characterize the different sources. The comparison between the XRF results on obsidian powders and the nondestructive XRF on obsidian integral fragments (concentration ratios and intensity ratios of the selected trace elements Nb, Y, Zr, Rb, and Sr) was carried out with very good results.

Fig. 18.2 Some obsidian splinters, detached from the geological obsidian sources samples. They are similar in shape to workshop refuse found in archaeological excavations

18

Provenance of Italian and Central European Archaeological Obsidians by. . .

511

Fig. 18.3 Comparison between the destructive and nondestructive XRF methodologies. a) Nb/Zr ppm vs. Nb/Sr ppm on powders; b) Nb/Zr vs. Nb/Sr (X-ray intensity ratios) on the entire fragments Mediterranean area. Acronyms: PT, Pantelleria; PL, Palmarola; LP, Lipari; SA, SB1, SB2, and SC, Sardinia – M. Arci; YA, Gyali; ML, Melos

In fact, for the Mediterranean area, the Nb/Zr vs. Nb/Sr plots (Fig. 18.3) allow to compare the trace element concentration ratios (ppm) obtained from traditional XRF method on the powdered obsidians (Fig. 18.3a), with the intensity ratios obtained by the nondestructive method on the entire fragments (Fig. 18.3b). The chemical composition of the geological obsidians of the Mediterranean area, Pantelleria, Palmarola, Lipari, Sardinia – Monte Arci SA, SB1, SB2, and SC – and those of the Greek islands of Melos and Gyali are all separated, without any overlap between the obsidian source groups, which are well-defined and perfectly comparable using both the above described analytical methods [21]. Figure 18.4 includes obsidian sources from the Carpathians and the Mediterranean together, aimed to verifying that they do not overlap. Indeed, as shown in Nb/Zr vs. Nb/Sr and in Rb/Sr vs. Zr/Y diagrams (Fig. 18.4a and c, respectively), the Carpathian obsidian sources are undoubtedly well separated from the other Mediterranean sources, although very close to Milos and to SC and SB1 Sardinian obsidian sources. Figure 18.4b and d represents the enlarged detail of Fig. 18.4a and c, which confirms and highlights more clearly that the sources of the Carpathians are very well discriminated from the sources of the Mediterranean [7]. These results confirm that the intensity ratios of only few trace element (Rb, Zr, Sr, Y, and Nb) obtained by nondestructive XRF method are more than sufficient to discriminate all the obsidian sources of the considered areas (Mediterranean and Carpathian Mountains), as with the traditional XRF method. Other trace elements too, such as Ba, Ti, Ce, and La, can give excellent results very useful for the same purpose.

18.5

Results and Concluding Remarks

The application of nondestructive WDXRF method to the obsidian fragments of 45 archaeological sites from Italy and Sardinia-Corsica and 7 from Central EuropeCarpathian Mountains allowed us to attribute the provenance for 1406 fragments out

512

A. M. De Francesco et al.

Fig. 18.4 Nb/Zr vs. Nb/Sr plot (a) and Rb/Sr vs. Zr/Y plot (c) intensity ratios of the Carpathian obsidian sources together with the Mediterranean obsidian sources; no overlap with each other subsists. Figure b and d is magnifications of the two previous figures. Acronyms: C1, Carpathian 1 – Slovakia; C2 Carpathian 2 T – Tolcsva; MAD, Carpathian 2E. The other acronyms as in Fig. 18.3

of 1453 archaeological obsidians. The results are summarized in Tables 18.1 and 18.2. The provenance of archaeological artifacts was determined by comparing the intensity ratios obtained on the obsidian splinters of known origin (geological sources) and those on artifacts from archaeological sites, analyzed simultaneously. Figure 18.5 illustrates graphically, as an example, the results of the analyses carried out on 64 obsidian fragments from the Neolithic archaeological site of Catignano, in the Abruzzo region (Italy), compared with all the possible Mediterranean and Carpathian obsidian sources. The Nb/Zr vs. Nb/Sr X-ray intensity ratios plot (Fig. 18.5a) refers to the nondestructive WDXRF analysis of obsidian fragments of Catignano archaeological site, represented by cross symbols. The overlap or the clear proximity of the archaeological samples with those of the obsidian sources determines the attribution to a geological source. As clearly shown in Fig. 18.5a, most of the archaeological obsidians (56 as shown in Table 18.2) come from the Lipari island, 6 from Palmarola island, and only 2 undefined provenances (Table 18.2). The triangular plot Nb/Sr – Nb/Y – Rb/Zr in Fig. 18.5b further confirms the attributed obsidian sources, more clearly separated also graphically from almost all the others. In conclusion, 1453 archaeological obsidians were analyzed by nondestructive XRF method, and 97% of them were assigned to their obsidian source. The provenance

Number of investigated archaeological sites 7 29 6 3 45

Geographical area Northern Italy Central Italy Southern Italy Sardinia-Corsica Total

Analyzed archaeological obsidian 54 1.163 61 120 1.398

Assigned to a source 50 1.125 61 120 1.356

Table 18.1 Obsidian fragments provenance of about 45 archaeological sites from Italy, Sardinia, and Corsica Undetermined 4 38 0 0 42

% of unassigned 7% 3% 0% 0% 3%

18 Provenance of Italian and Central European Archaeological Obsidians by. . . 513

514

A. M. De Francesco et al.

Table 18.2 Obsidian fragments provenance of seven archaeological sites from Central EuropeCarpathian Mountains [7]. In this table due to the small numbers, the name of the sites was also added

Site Iclod Limba Vararia Miercurea Sibiului Pestera Ungaresca Seimi Caramidarie Suplacu de Barcau Taga Total

Overall samples 11 2

Carpathian 1 Slovakia 10 1

Carpathian 2T Tolcsva 0 0

Mad 0 0

Undetermined 1 1

% of unassigned 9% 50%

1

1

0

0

0

0%

14

12

1

0

1

7%

11

11

0

0

0

0%

6

6

0

0

0

0%

10 55

8 49

0 1

0 0

2 5

20% 9%

Fig. 18.5 Examples of the application of the nondestructive XRF methodology on the obsidian fragments of Catignano, an Italian Neolithic site in the Abruzzo region. Acronyms as in Fig. 18.4. Cross symbols represent the archaeological obsidian samples

attribution of 1356 out of 1398 obsidian fragments from 45 Italian and Corsican archaeological sites was successful, as shown in Table 18.1. For the 7 Carpathian archaeological sites, 50 out of 55 obsidian fragments were attributed to an obsidian source (Table 18.2). The best results were achieved for the Italian archaeological sites (southern Italy and Sardinia-Corsica), for which all the obsidian fragments were assigned to the obsidian sources. Tables 18.1 and 18.2 also show the number of archaeological obsidians not assigned to an obsidian source together with their percentage. The percentage of the unassigned archaeological samples is very low and may be due to extremely small fragment size (less than 1 cm in diameter). The small number

18

Provenance of Italian and Central European Archaeological Obsidians by. . .

515

(about 3%) of unassigned archaeological obsidians represents a very good response to the application of a nondestructive analytical method. As a conclusion, the nondestructive WDXRF analytical method allows to trace the origin of archaeological obsidian in the studied areas, with precision and reliability similar to more sophisticated and expensive methods, reducing cost and analysis time.

References 1. Acquafredda P, Adriani T, Lorenzoni S, Zanettin E (1996) Proposal of a non-destructive analytical method by SEM-EDS to discriminate Mediterranean obsidian sources. Adv Clay Miner:269–271 2. Acquafredda P, Adriani T, Lorenzoni S, Zanettin E (1999) Chemical characterization of obsidian from different Mediterranean sources by non-destructive SEM-EDS analytical method. J Archaeol Sci 26:315–325 3. Barca D, De Francesco AM, Crisci GM (2007) Application of laser ablation ICPeMS for characterization of obsidian fragments from peri-Tyrrhenian area. J Cult Herit 8:141–150 4. Barca D, De Francesco AM, Crisci GM, Tozzi C (2008) Provenance of obsidian artifacts from site of Colle Cera, Italy, by LA-ICP-MS method. Periodico di Mineralogia 77:41–52 5. Bellot-Gurlet L, Le Bourdonnec F, Popeau G, Dubernet S (2004) Raman micro-spectroscopy of western Mediterranean obsidian glass: one step towards provenance study? J Raman Spectrosc 35:671–677 6. Bellot-Gurlet L, Popeau G, Salomon J, Calligaro T, Moignard B, Barrat JA, Pichon L (2005) Obsidian provenance studies in archaeology: a comparison between PIXE and ICP. Nucl Inst Methods Phys Res B 240:583–588 7. Biagi P, De Francesco AM, Bocci M (2007) New data on the archaeological obsidian from the middle-late Neolithic and chalcolithic sites of the Banat and Transylvania. In: Kozowski JK, Raczky P (eds) The Lengyel, polgar and related cultures in the middle/late neolithic in Central Europe. Polska Akademia Umiejetnosci, Krakow, pp 309–326 8. Bigazzi G, Yegingil Z, Ercan T, Oddone M, Ozdogan M (1993) Fission track dating obsidian in Central and Northern Anatolia. Bull Volcanol 55:588–595 9. Birò KT (1984) Distribution of obsidian from the Carpathian sources on Central European palaeolithic and mesolithic sites. Acta Archaeologica Carpathica 23:5–42 10. Birò KT (2004) Carpathian obsidians: myth and reality. In: Proceedings of the 34th ISA Institución “Fernando El Católico (CSIC) Excma”. Diputación De Zaragoza, Electronic Publication, pp 267–277 11. Birò TK, Pozsgai I, Vlader A (1986) Electron beam microanalyses of obsidian samples from geological and archaeological sites. Acta Archaeologica Academiae Scientiarum Hungaricae 38:257–258 12. Birò KT, Bigazzi G, Oddone M (2000) Instrumental analysis I. The Carpathian sources of raw material for obsidian tool-making. In: Dobosi VT (ed) Bodrogkeresztur-Henye. (NE-Hungary) Upper Palaeolithic site. Magyar Nemzeti Muzeum, Budapest, pp 221–240 13. Cann JR, Renfrew C (1964) The characterisation of obsidian and its application to the Mediterranean region. Proc Prehistoric Soc 30:111–133 14. Crisci GM, Ricq-De Bouard M, Lanzafame U, De Francesco AM (1994) Nouvelle méthode d'analyse et provenance de l’ensemble des obsidiennes neolithiques du Midi de la France. Gallia Préistoire 36:299–327 15. De Francesco AM, Crisci GM (1999) Provenienza delle ossidiane dei siti archeologici di Pianosa (Arcipelago Toscano) e Lumaca (Corsica) con il metodo non distruttivo in Fluorescenza X. In: Tozzi C, Weiss MC (eds) Il primo popolamento Olocenico dell’area corso-toscana; Edizioni. ETS, Pisa, pp 253–258

516

A. M. De Francesco et al.

16. De Francesco AM, Crisci GM (2003) L’ossidiana. In: Tozzi C, Zamagni B (eds) Gli scavi nel villaggio neolitico di Catignano (1971–1980). Origines. Studi e materiali pubblicati a cura dell'Istituto Italiano di Preistoria e Protostoria, Firenze, pp 239–240 17. De Francesco AM, Crisci GM, Bocci M, Lanzafame U (2002) Provenienza delle ossidiane archeologiche di alcuni siti neolitici italiani. PLINIUS, supplemento italiano all. Eur J Mineral 28:137–138 18. De Francesco AM, Bocci M, Crisci GM (2004) Provenienza dal Monte Arci per le ossidiane di alcuni siti archeologi dell’Italia Centrale e della Francia Meridionale. In: Proceeding of 2 Convegno Internazionale L’ossidiana del Monte Arci nel Mediterraneo. Pau (Cagliari) 28–30 Novembre 2003. Ed. AV (CA), pp 303–309 19. De Francesco AM, Bocci M, Crisci GM, Martini F, Tozzi C, Radi G, Sarti L, Cuda MT (2006) Applicazione della metodologia analitica non distruttiva in Fluorescenza X per la determinazione della provenienza delle ossidiane archeologiche del progetto “Materie Prime” dell’I.I.P.P. Atti della XXXIX Riunione Scientifica dell’Istituto di Storia e Protostoria 1:531– 548 20. De Francesco AM, Duran V, Bloise A, Neme G (2006) Caracterización y procedencia de obsidianas de sitios arqueológicos del área natural protegida Laguna del Diamante (Mendoza, Argentina) con metodología no destructiva por fluorescencia de rayos (XRF). Anales de Arqueología y Etnología 61:53–67 21. De Francesco AM, Crisci GM, Bocci M (2008) Non-destructive analytical method by XRF for determination of provenance of archaeological obsidians from the Mediterranean area. A comparison with traditional XRF method. Archaeometry 50(2):337–350 22. De Francesco AM, Bocci M, Crisci GM (2011) Non-destructive applications of wavelength XRF in obsidian studies. In: Shackley MS (ed) X-Ray Fluorescence Spectrometry in Geoarchaeology. Springer North America, New York, pp 81–107 23. De Francesco AM, Bocci M, Crisci GM, Francaviglia V (2012) Obsidian provenance at the several Italian and Corsican archaeological sites using the non-destructive X-ray Fluorescence method. In: Liritzis I, Stevenson CM (eds) Obsidian and ancient manufactured glasses – part II Obsidian Glass Provenance. University of New Mexico Press, Albuquerque, pp 115–129 24. De Francesco AM, Barca D, Bocci M, Cortegoso V, Barberena R, Yebra L, Duran V (2018) Provenance of obsidian artifacts from the natural protected area Laguna del Diamante (Mendoza, Argentina) and upper Maipo valley (Chile) by LA-ICP-MS method. Quat Int 468:134– 140, Part: A 25. Durán V, De Francesco AM, Neme G, Cortegoso V, Cornejo L, Bocci M (2012) Caracterización y procedencia de obsidianas de sitios arqueológicos del Centro Oeste de Argentina y Centro de Chile con metodología no destructiva por Fluorescencia de Rayos (XRF). Intersecciones en Antropología 13(2):423–437 26. Francaviglia V (1984) Characterization of Mediterranean obsidian sources by classical petrochemical methods. Preistoria Alpina – Museo Tridentino di Scienze Naturali di Trento (Italy) 20:311–332 27. Francaviglia V (1988) Ancient obsidian sources on Pantelleria (Italy). J Archaeol Sci 15:109– 122 28. Glascock MD, Neff H (2003) Neutron activation analysis and provenance research in archaeology. Meas Sci Technol 14:1516–1526 29. Gratuze B (1999) Obsidian characterization by laser ablation ICP-MS and its application to prehistoric trade in the Mediterranean and the near east: sources and distribution of obsidian within the Aegean and Anatolia. J Archaeol Sci 26:869–881 30. Hallam BR, Warren SE, Renfrew C (1976) Obsidian in the western Mediterranean: characterization by neutron activation analysis and optical emission spectroscopy. Proc Prehistoric Soc 42:85–110 31. Kayani PI, McDonnel G (1996) An assessment of back-scattered electron petrography as a method for distinguishing Mediterranean obsidians. Archaeometry 38:43–58

18

Provenance of Italian and Central European Archaeological Obsidians by. . .

517

32. Machey MP, Warren SE (1983) The identification of obsidian source in the M. Arci region of Sardinia. In: Aspinall A, Warren SE (eds) Proceedings of the 22nd symposium on Archaeometry. University of Bradford, Bradford, pp 420–431 33. Oddone M, Mfirton P, Bigazzi G, Biró KT (1999) Chemical characterisations of Carpathian obsidian sources by instrumental and epithermal neutron activation analysis. J Radioanal Nucl Chem 240(1):147–153 34. Renfrew C, Cann JR, Dixon JE (1965) Obsidian in the Aegean. Annu Br School Athens 60: 225–247 35. Rosania CN, Boulanger MT, Biró KT, Ryzhov S, Trnka G, Glascock MD (2008) Revisiting Carpathian obsidian. Antiquity 82(318). Project Gallery. http://antiquity.ac.uk/projgall/rosania/ 36. Shackley MS (ed) (1998) Archaeological Obsidian studies: method and theory. Kluwer Academic/Plenum Press, New York and London 37. Shackley MS (1998) Gamma rays, X-rays, and stone tools: some current advances in archaeological geochemistry. J Archaeol Sci 25:259–270 38. Shackley MS (2002) Precision versus accuracy in the XRF analysis of archaeological obsidian: some lesson for archaeometry and archaeology. In: Jerem E, Biro KT (eds) Proceedings of the 31st symposium on Archaeometry, Budapest, Hungary. BAR International series 1043(II). BAR International, Oxford, pp 805–810 39. Shackley MS (2005) Obsidian. Geology and archaeology in the North American Southwest. University of Arizona Press, Tucson 40. Stewart SJ, Cernicchiaro G, Scorzelli RB, Poupeau G, Acquafredda P, De Francesco AM (2003) Magnetic properties and 57 Fe Mossbauer spectroscopy of Mediterranean prehistoric obsidians for provenance studies. J Non-Cryst Solids 323:188–192 41. Summerhayes G, Bird JR, Fullagar R, Gosden C, Specht J, Torrence R (1998) Application of PIXE-PIGME to obsidian characterization on west New Britain, Papua New Guinea. In: Shackley MS (ed) Archaeological Obsidian studies: method and theory. Plenum Press, New York/London, pp 129–158 42. Tykot RH (1997) Characterization of the Monte Arci (Sardinia) obsidian sources. J Archaeol Sci 24:467–479 43. Tykot RH (2017) A decade of portable (Hand-Held) X-ray fluorescence spectrometer analysis of obsidian in the mediterranean: many advantages and few limitations. MRS Adv. https://doi. org/10.1557/adv.2017.148 44. Tykot RH, Young SMM (1996) Archaeological applications of inductively coupled plasmamass spectrometry. In: Orna MV (ed) ACS symposium series 625, Archaeological chemistry. Washington, DC, pp 116–130 45. Tykot RH, Setzer T, Glascock MD, Speakman RJ (2005) Identification and characterization of the Obsidian sources on the Island of Palmarola, Italy. Geoarchaeol Bioarchaeol Stud 3:107– 111 46. Williams-Thorpe O (1995) Obsidian in the Mediterranean and the near east: a provenancing success story. Archaeometry 37:217–248 47. Williams-Thorpe O, Warren SE, Courtin J (1984) The distribution and sources of archaeological obsidian from Southern France. J Archaeol Sci 11:135–146

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and Characteristics for In Situ Cultural Heritage Studies

19

Carlos Roberto Appoloni, Fabio Lopes, Paulo Sergio Parreira, Tiago Dutra Galva˜o, Fabio Luiz Melquiades, Renato Akio Ikeoka, and Eduardo Inocente Jussiani

Contents 19.1 19.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . pXRF Studies of Churches Mural Paintings: Characterization and Conservation Procedures Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.3 pXRF Studies of Gold and Silver Historical and Archaeological Objects: Elemental Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.3.1 Case Study: Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.3.2 Case Study: Small Bird . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.4 pXRF Measurement of Elemental Composition and Multilayer Thickness by Kα/Kβ Ratios of Metals and Pigments of Objects in the Cultural Heritage . . . . . . . . . . . . . . . . . . . . 19.4.1 Six Little Golden Heads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.4.2 Differential Attenuation of Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.5 pXRF, Portable Raman Spectroscopy, and TXRF Paintings Examination . . . . . . . . . . . . . 19.5.1 Case Study: “Moema” Painting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6 pXRF Studies of Archaeological Ceramics and Obsidians: Characterization and Provenance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6.1 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6.2 Results and Discussions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.6.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19.7 pXRF and Portable Raman Spectroscopy for In Situ Rock Art Analysis . . . . . . . . . . . . . . 19.8 Gold and Silver Coins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

520 520 522 523 523 524 527 528 530 531 536 537 541 546 547 551 554

C. R. Appoloni (*) Physics Department/CCE, State University of Londrina, Londrina, Brazil e-mail: [email protected] F. Lopes · P. S. Parreira · T. D. Galvão · F. L. Melquiades · R. A. Ikeoka · E. I. Jussiani State University of Londrina, Londrina, Brazil e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_19

519

520

C. R. Appoloni et al.

Abstract

Laboratory portable X-ray fluorescence (pXRF) systems are widely applied for cultural heritage studies. This chapter will deal with some applications such as studies of churches mural paintings, studies of gold and silver historical and archaeological objects, measurement of elemental composition and multilayer thickness by Kα/Kβ ratios of metals and pigments of objects in the cultural heritage, paintings examination, in situ rock art analysis, and studies of archaeological ceramics and obsidians. Laboratory systems are very versatile since their collimation and positioning of the excitation/detection system can be adapted to many different geometries of the samples, which can be a small coin or a statue of human size, a large painting, or a complex metal object composed of many different parts, as well as change the X-ray tube for measurements with different tube anodes.

19.1

Introduction

Portable X-ray fluorescence (pXRF) instruments have gained considerable acceptance within the archaeological community [1]. Moreover, it has gradually become a routine step in several restauration and authentication processes applied in different museums around the world [2]. The attractiveness in pXRF is a simple, fast, nondestructive, and multielement technique, no sample alteration is caused and when necessary safe sample preparation without chemical waste may be performed. However, pXRF disadvantages are the small penetration of the X-rays, typically tens of micrometers. pXRF do not detect light elements (under Z ¼ 12) neither access chemical speciation. The basic principle of X-ray fluorescence is based in the photoelectric effect. It is built on the fact that a sample, when irradiated with energetic X-rays, has a given probability of emitting characteristic X-rays [3]. The data collection enables qualitative identification of the elements as well as its quantitative concentration or layer thickness determination, only if evaluated with appropriate quantification procedures [4, 5]. Although quantification analytical methods are well established in the scientific community, many of the archaeologists and restaurateurs’ queries are solved with qualitative or quali-quantitative analysis. This chapter addresses pXRF applications in cultural heritage studies. In the sequence several cases of study which deal with different kinds of objects and different questions to be answered will be detailed.

19.2

pXRF Studies of Churches Mural Paintings: Characterization and Conservation Procedures Monitoring

With the aim for helpness into the evaluation of the problems associated with the restoration of afrescos paintings [6–8] in church walls, like water infiltrations, efflorescence of mineral salts, and the composition of the pigments, it was used a

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

521

portable EDXRF system [9] developed by GFNA (Applied Nuclear Physics Group) from the State University of Londrina, in a scientific cooperation with two restorers, during the restoration process of Immaculate Conception Parish Church, of São Paulo City, and Immaculate Conception Basilica, of Botafogo in Rio de Janeiro City. Some paintings of the São Paulo Church have shown a whitish cover due to the efflorescence of salts brought up with water infiltration by capillarity throughout the walls, when compared with painting without infiltration, as could be seen in Fig. 19.1a, b. During the process of cleaning the whitish pictorial layers, measurements were taken to obtain the most effective chemical quellants in the removal of the whitish layer and dirtiness because the pollutants present in the air [10, 11]. It was identified the presence of the macroelements S, K, Ca, Ti, Cr, Mn, Fe, Co, and Zn which were associated with the following pigments: zinc white (ZnO); titanium dioxide (TiO2); brown pigments like brown ocher (Fe2O3.nH2O), Umbria pigments, or Van Dyke brown pigment; yellow ocher as a mix of brown ocher and white pigments; the green pigments as chromium dioxide (Cr2O3) or veridian (Cr2O3.2H2O); and the blue pigment which was associated with a French Ultramarine (Na8-10Al6Si6O24)S2-4 or phthalocyanine (C32H18N8); these last ones are organic pigments, not detected by pXRF. All these pigments were identified by the restorer with the identified XRF elements and the historical recording about the Church building allowing a more reliable cleaning and restoration process with the original paint. The Church of the Immaculate Conception Basilica is a neo-Gothic church style of the late nineteenth century and is located in Botafogo-RJ. It was declared as Rio

Fig. 19.1 Medallion before (a) and after (b) the cleaning process

522

C. R. Appoloni et al.

Fig. 19.2 Wall painting of the altar showing the different colors and part of the EDXRF portable system (DSCF3216)

de Janeiro’s cultural heritage, and the objective of this work was to characterize the pigments used, both in the original painting, Fig. 19.2, and in the interventions, Fig. 19.3, which occurred in 1979 and another in an indeterminate time, with the purpose of assisting the restorer in the process of recovering the original painting as well as the restoration process itself. Measurements were made in regions of the restoring wall as well as in regions of the walls that were not under restoration, as reference sites [12]. Figure 19.2 shows a part of the wall painting of the altar under restoration which shows regions with different shades of red, blue, gold, and yellow colors. The macroelements which characterize the different pigments with shades of white, red, blue, gold, and yellow colors were, basically, S, Ca, Ti, Fe, Cu, Sn, Hg, Au, and Pb. The presence of the elements Au and Cu in the golden lace is really a thin film of gold applied in all these decorations in the altar walls. The elements Hg and S were identified in the red colors showing the vermilion pigment (HgS), while the green pigments and its nuances are due to the mix of Prussian blue (Fe4[Fe(CN)6]3) and chrome yellow (PbCrO4). The Ca and Pb elements were identified in all the regions measured, indicating for the restorer the technique named tempera painting, and used as a base to receive the other pigments. The presence of Ti in the repainting regions clearly indicates the use of modern pigments.

19.3

pXRF Studies of Gold and Silver Historical and Archaeological Objects: Elemental Characterization

Two case studies are presented as examples of this kind of application.

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

523

Fig. 19.3 Wall of repainting behind the pulpit showing part of the EDXRF portable system (DSCF3243)

19.3.1 Case Study: Board It was analyzed a piece in plate format, probably used as a pectoral adornment, shown in Fig. 19.4, belonging to the collection of the Municipal Museum of Piura, Peru [13]. Archaeologists believe that this piece is supposedly belonging to the Chavín culture due to the characteristics of form and region where it was found. The result obtained is shown in Table 19.1, and the resulting spectrum of analysis is shown in Fig. 19.5. The results indicated an alloy with high concentration of gold, evidencing the great ability in the Chavín culture to obtain gold with great purity. It was not necessary to perform the calculations of self-attenuation of gold for this piece, since the concentrations of silver and copper are small; therefore the piece could not be made of tumbaga or copper plated with gold.

19.3.2 Case Study: Small Bird A gold ornament representing a small bird with a pendant in the beak, belonging to the Vicús culture, of the Municipal Museum of Piura, Peru, was analyzed in one region [13]. A photo is shown in Fig. 19.6.

524

C. R. Appoloni et al.

Fig. 19.4 Gold plate of the Chavín culture. Image of the collection of the Municipal Museum of Piura, Peru

Table 19.1 Concentration of gold, silver, and copper for the plate. Culture Chavín. Municipal Museum of Piura, Peru Gold (%) 92.5  0.9a

Board a

Silver (%) 7.1  0.5

Copper (%) 0.40  0.07

Standard deviation

The analysis of the elemental concentration for the small bird in gold is presented in Table 19.2. Due to the morphological characteristics and the good conservation status of the part, the result of the pXRF analysis indicates that the small bird was made in an AuAg-Cu alloy, with gold as the major element. The spectrum illustrating the analyzed point is shown in Fig. 19.7.

19.4

pXRF Measurement of Elemental Composition and Multilayer Thickness by Kα/Kβ Ratios of Metals and Pigments of Objects in the Cultural Heritage

When thin layers of elements (sheets) are superimposed, such as layers of pigments in paint, thin sheets of gold, silver, zinc, decorative pigments on ceramics, etc., the thicknesses of these layers can be determined by calculating the different

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

525

Fig. 19.5 Spectrum of the plate. Municipal Museum of Piura

Fig. 19.6 Small bird in gold. Collection of the Municipal Museum of Piura, Peru. Culture Vicús

Table 19.2 Results of analyses of gold, silver, and copper concentrations of the small bird in gold. Collection of the Municipal Museum of Piura, Peru Small bird Point 1 a

Gold (%) 89.1  1.2a

Silver (%) 7.4  0.5

Copper (%) 3.5  0.4

Standard deviation

attenuations for the lines K and L. An illustrative scheme is shown in Fig. 19.8. Thus, for example, in the case of a base of an element “a” such as copper covered with a thin layer of another element “b” such as gold (Fig. 19.8), the ratio (Kα/Kβ) of the element “a” present in the innermost layer depends on the thickness and

526

C. R. Appoloni et al.

Fig. 19.7 Spectrum of the small bird in gold. Collection Municipal Museum of Piura, Peru. Culture Vicús

composition of the element “b.” The same is true for the ratios between the lines (Lα/Lβ) [4, 14]. These relationships provide the basis for the analysis of pictorial layers on canvases, gold, silver, and other artifacts that contain more than one element. It is also an excellent methodology for the determination of thin film thicknesses [15]. Thus when X-rays of the layer K or L of an element cross a layer of another element, the ratio between the lines is changed differently, due to the energy difference between the lines Kα, Kβ and Lα, Lβ. This attenuation is called differential attenuation and given by Eqs. 19.1 and 19.2 which in this work will be indicated by the letter R, so:


where:

R¼

 Kα eðμ1 μ2 Þρd Kβ 1

ð19:1Þ

R¼


 Lα eðμ1 μ2 Þρd Lβ 1

ð19:2Þ

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

527

Fig. 19.8 Illustrative scheme representing the attenuation of copper by gold

•

 

Kα Kβ 1

  and

Lα Lβ 1

are the ratios for the innermost metal, of infinite thickness.

• μ1 is the linear attenuation coefficient in cm1 of the material lying on the surface for the energy radiation Kα of the innermost element. • μ2 is the linear attenuation coefficient in cm1 of the material lying on the surface for the energy radiation Kβ of the innermost element. • ρ is the density in g/cm3 of the material. • d is the thickness in centimeters of the material on the surface.   The ratios

Kα Kβ 1

  and

Lα Lβ 1

represent the ratio of the material due to self-listing

and are determined experimentally and  tabulated [16, 17]. For example, for the copper and silver elements, the ratio Kα Kβ 1 is 5.1 and 6.9 respectively; and for the   gold element, the ratio Lα is approximately 1 [18, 19]. Lβ 1

19.4.1 Six Little Golden Heads A piece forming a set of six (06) small golden heads, shown in Fig. 19.9 belonging to the collection of the Enrico Poli Museum in Lima, Peru, was also studied [13]. The state of preservation of the piece is excellent, and although visually all the small heads have a golden appearance, the analysis of the X-ray spectra indicated an alloy highly rich in copper, with presence of gold and zinc in a lower concentration. The results are shown in Table 19.3.

528

C. R. Appoloni et al.

Fig. 19.9 Six little golden heads. Chavín culture. Image of the collection of the Enrico Poli Museum, Lima

Table 19.3 Results of the analyses of the six small golden heads. Chavín culture, Enrico Poli Museum, Lima

Golden heads Head 1 Head 2 Head 3 Head 4 Head 5 Head 6 a

Gold (%) 11.6  0.6a 8.1  0.3 4.8  0.2 5.7  0.4 9.1  0.4 10.0  0.5

Copper (%) 86.2  1.3 89.4  1.1 93.1  1.0 91.9  1.3 88.8  1.4 87.9  1.2

Zinc (%) 2.19  0.02 2.59  0.02 2.15  0.01 2.42  0.03 2.10  0.01 2.13  0.02

Standard deviation

Analyzing the X-ray spectra in conjunction with the results of Table 19.3, it can be concluded that the six small gold heads were produced from a very pure copper alloy (tumbaga) with the presence of only zinc traces with an average concentration of 2.26%. The silver identified in the spectra comes from the X-ray tube. Two spectra are shown in Figs. 19.10 and 19.11. Although the set of heads has a gold coloration, it was evident from the analyses presented that the piece was produced from the tumbaga alloy, due to the high concentration of copper. Thus, the differential attenuations, as well as the reasons, were calculated by the methodology described previously.

19.4.2 Differential Attenuation of Copper The calculated values for differential attenuation for copper Cu Kα Kβ at all points analyzed are shown in Table 19.4. Calculation of these ratios indicated different depths for gold in all six small heads. This fact demonstrates that the heads are composed of tumbaga in copper and that the average equivalent thickness of gold with this methodology was 1.25  0.39 μm.

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

529

Fig. 19.10 Spectrum of the head 3. Culture Chavín, Enrico Poli Museum, Lima

Fig. 19.11 Spectrum of the head 6. Chavín culture, Enrico Poli Museum, Lima

Analyses of the results presented in Tables 19.3 and 19.4 for the ratios demonstrate that the set of six small gold heads was produced with a copper tumbaga alloy. The mean equivalent thickness for gold, calculated by self-attenuation for gold and differential attenuation for copper, was 1.31  0.54 μm.

530

C. R. Appoloni et al.

Table 19.4 Calculation of differential attenuation of copper for the six small golden heads. Chavín culture, Enrico Poli Museum, Lima

Six small golden heads Head 1 Head 2 Head 3 Head 4 Head 5 Head 6 a

19.5

Measured ratio   Copper Kα Kβ 4.96  0.06a 4.70  0.05 5.07  0.06 5.23  0.06 5.08  0.06 5.03  0.06

pXRF Thickness Gold (μm) 1.10  0.12 1.70  0.17 0.87  0.19 0.52  0.10 0.83  0.10 0.97  0.10

Standard deviation

pXRF, Portable Raman Spectroscopy, and TXRF Paintings Examination

Two categories of methods are normally used to identify the pigments employed in a painting: (a) those that measure the elemental chemical composition of the region of interest in the painting, and, depending on the elements observed, the pigments are inferred based on knowledge of the chemical composition of pigments employed at that time [20], and (b) those that measure the molecular “signature” of the materials in the region of interest and comparing to a database of molecular “signatures” of pigments in order to identify them [21]. There is a great variety of methodologies within the two categories, which can be divided into portable methods that can be performed in situ, at the museum or other location where the work is located, and those performed in laboratories, with greater precision and at greater cost, that require the artwork to be transported to them, which is not always possible or desirable. The most often used method for analysis of the elemental chemical composition is X-ray fluorescence, due to its extensive application, low cost, and its noninvasive, nondestructive, and simultaneous multielemental measurement that does not require sampling. This method is widely employed for the analysis of inorganic materials and some inorganic loaders signatures of organic pigments, as lacquers. Its portable version was made available relatively recently and is known by the acronym pXRF (portable X-ray fluorescence) [22–24]. Compact equipment was also recently made available in another modality within the X-ray fluorescence methodology, called total reflection X-ray fluorescence (TXRF), which requires removing only a few nanograms of sample to perform the measurement. This sampling is completely imperceptible in the painting [25]. When X-ray fluorescence is not enough to resolve doubts between the possible inorganic pigments and/or there are organic materials to be identified (pigments, overpainting, varnish, binding agents), the Raman spectroscopy method is one of the most often recommended. It is very good for laboratory analyses, but with sampling based on at least milligrams (depending on the equipment), and has also recently been converted into portable versions, which are less accurate, but allow for measurements in situ and without any sampling [26].

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

531

19.5.1 Case Study: “Moema” Painting “Moema” was performed by the famous Brazilian painter Victor Meirelles (1832– 1903) at 1866, has 1.90 by 1.29 m, is one of the most important Brazilian paintings of the nineteenth century, and belongs to the collection of Assis Chateaubriand São Paulo Art Museum (MASP). The examination of the painting was performed using the pXRF methodology, with supplementary measurements by TXRF and portable Raman spectroscopy (PRS) [27]. Figure 19.16 shows the system employed. The portable X-ray fluorescence system used consists of a mechanical support specifically developed to accommodate a mini-X-ray tube (used to excite the sample) and a mini-detector of the characteristic X-rays (to obtain the energy spectra of the X-rays emitted by the sample), set at a 90 angle from each other [9]. The system has 3 degrees of freedom in relation to the sample: linear, rotational, and angular. The equipment used, owned by LFNA/UEL [9], consisted of a mini-X-ray tube with a silver anode (model 2022) and a FTC-200 high-voltage system manufactured by the company MOXTEK, Inc., with a 50 μm Ag filter at the end of the tube; a Si-PIN (171 eV of FWHM for the 5.9 keV line) semiconductor detector, model XR100-CR, high-voltage source and amplifier model PX2CR, multichannel model MCA 8000ª, 2 mm diameter Ag collimator at the detector entry point, and PMCA spectrum acquisition software, all modules made by the company Amptek Inc.; and a notebook for data acquisition and analysis. The measurement conditions were 28 kV voltage applied to the tube, 5microA current on the tube, and excitation/detection time of 300 s (live time). All spectra were analyzed using the software QXAS-AXIL by the IAEA (International Atomic Energy Agency). Based on the net areas and deviations of the peaks supplied by this program, the presence of the elements listed in the previous items had the following criteria: net area of the peak corresponding to an element greater than three times the standard deviation of it and net area greater than three times the square root of background counts under the peak (lowest detection limit). The total reflection X-ray fluorescence (TXRF) methodology is a variant of EDXRF in which the incidence of the beam for exciting the sample is at the critical angle of Snell’s law. Due to its position, the resulting beam does not interact with the support, but penetrates the entire thin film formed by the deposition of the sample, both in the sense of incidence and emergence, thus making it more likely that it will excite the atoms that compose the sample. The TXRF measurement was performed using the Q-tip methodology [25], meaning samples with a few dozen nanograms were removed by the restorer from a few points of the painting by the delicate abrasion of cotton swabs. The cotton swabs were taken for measurement at LFNA/UEL. The measurement of the samples was performed using a BRUKER AXS S2 PICOFOX TXRF system. The basis of the Raman spectroscopy technique is the excitation of the sample by a laser and detection of the observed Raman transition spectra (difference between the Rayleigh scattering and Raman Stokes scattering of the light). With this type of

532

C. R. Appoloni et al.

spectrum, it is possible to determine the molecular composition of the sample. Each peak corresponds to a vibration mode excited by a molecule present in the sample, and, often, one single molecule is capable of generating various peaks. Thus, the peaks of the spectrum characterize the molecular composition of the sample. Therefore, a Raman spectrometer has monochromatic (laser) light source within the visible or ultraviolet spectrum used to excite samples, a detector which identifies the Raman transitions and separates them according to the number of wavelengths within a spectrum, along with the electronics necessary to power the system. The portable Raman spectrometer used in the measurements was an Inspector model by DeltaNu, which contains a 785 nm laser with a maximum power of 120 mW, a resolution of 8 cm1, and spectral reach for Raman transitions of 200– 2000 cm1. Seventy-three regions of the painting were measured, of which 23 were measured before and after the cleaning and another 50 regions only after the cleaning. These regions were chosen by mutual agreement with the restorer. Figure 19.12 presents the painting with the location of the 73 points measured. The painting was made using a relatively restricted palette, using a large variety of iron oxide-based pigments, from yellow to brown, all of them with russet tones. The entire painting is dominated by reddish-ocher tones. The pXRF analysis allows identification of the chemical elements and, therefore, the key elements of the inorganic pigments or those organic pigments whose composition contains chemical elements with an atomic number higher than 14 (in the case of the measurement

Fig. 19.12 Photograph of the painting with the location of the points measured

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

533

system used in this work). The chemical elements identified in the various original regions of the work were Ca, Fe, Cu, Hg, and Pb, indicating a palette composed of [28] yellow, red, and brown iron oxide ochers, Prussian blue, earth green, lead sulfide (galena) or lead oxide for black, verdigris, calcium carbonate (limestone/ chalk/calcite) or hydrated calcium sulfate (Gypsum), vermilion (mercury sulfide), and white lead (lead carbonate). All 73 regions measured displayed the X-rays characteristic to the Fe, Hg, and Pb, regardless of the color viewed. The Pb line is always very intense, from tens to hundreds of times greater than those of the other elements measured. Lighter regions showed more Pb than the others, indicating the use of white lead to lighten the mixture of pigments in the region. Red, or redder than average, regions contain greater quantities of Hg and/or Fe, indicating increased quantity of vermilion and/or red ocher. This leads to the hypothesis that the preparation layer(s) is(are) made using a white lead (lead carbonate) and vermilion (mercury sulfide) base and perhaps also mixed with ochers (hydrated iron oxides). A macrophotograph shows grains of pigment of a reddish tone in the preparation layer, another evidence to support this hypothesis. Only 18 of the 73 regions measured did not display quantifiable lines of Ca, or rather 25% of the regions. However, this does not discard the possibility of calcium carbonate (limestone/chalk/calcite) or hydrated calcium sulfate (gypsum) being used in the preparation layer since X-rays of Ca have low energy and may have been absorbed by the painting in regions where the quantity of Ca-based materials is small and/or there is a relatively dense layer of other pigments on top. Each color region was analyzed, and as an example, pink and red regions are characterized by the use of red ocher and vermilion, in addition to white lead. Figure 19.13 presents the spectra of three regions, Point 15, region of the dark red feather; Point 23, pink color of the sky; and Point 37, red feather on the belt, showing the variation of elements with the color. The quantity of red ocher and vermilion used in the pigment mixture varies depending on the tone desired by the artist. Nine regions characterized as prior restorations, given they presented elements other than those of the overall palette of the painting, were measured. Based on the elements presented, these regions may be divided into three groups, based on the date the pigments involved were introduced. Three regions, Points 16, 17, and 38, presented lines of Zn, Cr, and Ti, characterizing them as restorations performed in the twentieth century, due to the introduction of the titanium white pigment beginning in 1923. Another region, point 11, presented Zn and Ba, characteristic of the use of lithopone, introduced in the nineteenth century beginning in 1874, meaning this restoration was performed at least 8 years after the painting was made, given it was painted in 1866. The other regions, Points 18, 57, 58, 62, and 73, present only Zn, introduced in the market beginning in 1835, indicating that the restoration occurred in the period extending from after the painting was made in the nineteenth century until the start of the twentieth century. Samples were collected from eight points of the painting, some of which still contained an extremely thin layer of varnish, resulting in none of the pigment spectra. At other points it was possible to observe the spectrum of the extremely superficial layer of the painting after reduction of the varnish. Figure 19.14 presents

534

C. R. Appoloni et al.

Fig. 19.13 Spectra for X-rays (counts) versus energy (keV) of Point 15 (green line), dark red; Point 23, pink (violet line); and Point 37, red (blue line). The dark red feather (green line) is notably the one with the highest Hg peak, meaning it contains more vermilion pigment, and an intermediate value of Fe, that is, red ocher pigment. The red feather (blue line) contains much less Hg, vermilion, and more Fe, that is, red ocher, in the pigment composition. The pink region, violet line, is the one containing the least Fe, red ocher, and a little more Hg, vermilion, than the red feather, and more Pb, white lead, than the other two regions, in order to result in this sky-pink tone

the TXRF spectrum of a region near Point 71, the light green feather. Note that even in this superficial layer, there is a strong presence of the Pb and Hg lines. The presence of a K peak over a significant area corroborates the evidence of the use of earth green, which, in addition to Fe, also presents K in its composition. The portable Raman spectroscopy measurements were performed before the cleaning, that is, before removal of the varnish, in order to verify the potential of the methodology, even with the possible interference between the layer of varnish and the pigments. Points 20, 21, 22, and 23 were measured. The lines observed at Point 21, dark pink, are those corresponding to the HgS molecule, that is, the vermilion employed to provide the tone of the color in this region. Figure 19.15 presents the Raman spectrum for Point 23, pink. Considering two standard deviations in the measured value of the centroid of each line, among the lines observed, five of them are attributable to gamboge resin, indicating that the overpainting of the work was performed using this resin, whose use was widely disseminated at the time the painting was made. This is the only direct proof about the use of this material by the artist.

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

535

Fig. 19.14 TXRF spectrum for the region near to Point 71. The Pb and Hg peaks are evident, in addition to the presence of Fe, Ca, and K

Fig. 19.15 Raman spectrum for Point 23 performed at high power (51 mW). Lines to the right, in the region from 1259 to 1628 (cm1), can be attributed to the resin from the overpainting to a binding agent and to egg whites, as discussed in the text

536

C. R. Appoloni et al.

Table 19.5 Synthesis of the pigments indicated by the measurements performed

Color White White White Red Red Green

Green Yellow Brown Blue Blue

Name of pigment White Lead carbonate lead Calcite Calcium carbonate Gypsum Hydrated calcium sulfate Vermilion Mercury sulfide Red Hydrated iron III oxide + ocher clay + silica Earth Hydrated magnesium, green iron, and potassium aluminum silicate Verdigris Copper acetate Yellow Hydrated iron III oxide + ocher clay + silica Brown Hydrated iron oxide + ocher clay + silica Prussian Hydrated ferric blue hexacyanoferrate Verdigris Copper acetate

Chemical formula 2PbCO3.Pb(OH)2

Key element or indicator for XRF Pb

CaCO3 CaSO4.2H2O HgS Fe2O3.H2O

Ca Ca, S Hg, S Fe

Variations of K[(AlIII, FeIII)(FeII, MgII)], (AlSi3, Si4)O10(OH)2 Cu(C2H3O2)2.2Cu(OH)2 Fe2O3.H2O

Fe, K

Cu Fe

Fe2O3.H2O

Fe

Fe4[Fe(CN)6]3.14-16H2O

Fe

Cu(C2H3O2)2.2Cu(OH)2

Cu

On the other hand, four lines could be attributable to a binding agent similar to linseed oil, one of which belongs only to this type of oil, indicating that the binding agent of the pigments would have had the linseed oil in its composition, a fact already observed in other paintings by Victor Meirelles. Four lines could also have been attributed to egg whites, a material found in two other works by the same painter and used as varnish, but its strongest line at 1002 cm1 was not observed. Since a large part of the lines overlap within the resolution of the measurements, all three materials could be present. Considering all of the analyses performed, we can synthesize that the likely inorganic pigments used by the painter are those listed in Table 19.5. In relation to the organic materials, there is evidence of the use of gamboge resin and the binding agent linseed oil, as well as the possibility of the use of egg whites.

19.6

pXRF Studies of Archaeological Ceramics and Obsidians: Characterization and Provenance

The study of material culture is an important means for understanding the pre-colonial history of any country. Among all these material culture, ceramics are of great archaeological value, since they are extremely resistant to weather and the surrounding media conditions [29]. This also represents an integral part of a society, presenting sensitive singularities at a material level, which characterize the technological system and the way of life of the peoples studied [30].

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

537

It is also important to emphasize that the technological conditions of the ceramic inform the particularities of obtaining, handling, treating, and using the raw material, which encompass the relationships with the environment and the work object beyond the productive process, whereas the study of forms allows to know the functional character elements, such as production modes and means, and their correlation with aesthetic patterns. It is not enough to only obtain material testimonies from the past to the scientific knowledge of this field, but it is also necessary to understand and clarify the role of human groups from the artifacts, such as explaining the behavior and cultural change; symbolic, social, and technological aspects; as well as the processes for the formation of sites for both past and present peoples. In order to explain this complexity of tasks, a multidisciplinary study is pivotal, such as the case of archaeometry. Archaeometry is a physical-chemical research area studying problems related to cultural inheritance. It is based on the obtainment of information regarding the origin and history of the findings, material analysis related to the chemical structure and changes, as well as dating techniques [31]. The chemical characterization of such fragments may provide important information about the origin of the raw material, the quality of the coating, evidence of the occurrence of paintings, etc., which help the archeological studies about the ancient peoples. With the portable X-ray fluorescence (pXRF) technique, it is possible to make the chemical characterization of pottery fragments and obtain important information about the origin of the raw material and the quality of the revetment, among others, in a multielemental, simultaneous, fast, and nondestructive way [32–36]. The study of archaeological and cultural heritage artifacts by means of analytical techniques with portable equipment has become increasingly routine today. Various types of portable EDXRF equipment have been used in many different situations involving in situ analysis covering a wide range of geometries, detectors, current, and voltage applied in the X-ray tubes. Therefore, it’s extremely important that the geometry of the portable systems used in these studies be suitable for accurate acquisition and subsequent analysis of data. The pottery fragments obtained from Sambaqui do Bacanga were analyzed using a pXRF aiming to verify the existence of any type of different treatment on faces (concave and convex) related to the ceramic paste. The multi-varied analysis was employed to verify the similarity among the chemical elements in fragments with equal and different stratigraphies. In addition to the analysis of the pottery fragments, samples of obsidian were analyzed using two distinct pXRF systems.

19.6.1 Materials and Methods Pottery Fragments The pottery fragments were collected in an archeological site in Sambaqui do Bacanga, located in the Bacanga National Park in the island of São Luís – MA/Brazil.

538

C. R. Appoloni et al.

Table 19.6 Information about fragments Sector Exploratory trench Excavation area 1 Profile 1 Profile 2 Total

Quantity of fragments 34 6 24 4 68

Stratigraphic levels (cm) Surface up to 150 Surface Surface up to 140 30 up to 40

In the archeological excavations in Sambaqui do Bacanga, several types of material vestiges have been evidenced, like fire remains, human skeletons, remains of animal bones, fish remains, shells, mollusk shells, and ceramic objects. Ancient civilizations that inhabited that territory were characterized by being fishing populations, catchers, hunters, and ceramists, who settled down in the region around 6,600 years ago and remained inhabiting the place till the year 900 of our era. Stratigraphic samples of 68 pottery fragments were collected, deriving from the following sectors: exploratory trench, profile 1, excavation area 1, and profile 2, as shown in Table 19.6. Samples sizes varied from 3.5  2.5 cm to 11.0  6.5 cm.

Portable System of EDXRF in Pottery Fragments Analysis The portable system of EDXRF (PXRF-LFNA-02) utilized for irradiation/detection of the fragments consists of a 4W X-ray mini-tube with Ag anode and 50 μm Ag filter (MOXTEK, Inc.); a Si-PIN detector with preamplifier XR-100CR (Amptek Inc.), 221 eV FWHM for the 5.9 keV Mn line, 25 μm Be window, and thermoelectric cooling system by Peltier effect; PX2CR conjugated high tension source module and amplifier (Amptek Inc.); multichannel analyzer model MCA8000A (Amptek Inc.); an excitation/detection system positioning module with freedom degrees XYZ and rotation in relation to the analyzed sample; and a notebook for data acquisition and storage [9]. Measurement conditions were 28 kV and 5 μA in the X-ray mini-tube, and acquisition time was 500 s. In average, nine measurements were made for each fragment, three of which in the concave face, three in the convex face, and three in the ceramic paste. The analysis of the spectra was conducted by using the software Quantitative X-Ray Analysis (AXIL) [37], distributed by the International Atomic Energy Agency (IAEA). For the analyses of the data obtained with AXIL, only the intensities of the chemical elements whose net areas were threefold higher their standard deviations were considered. Obsidians This study’s main objective was to characterize the elemental chemistry of seven samples of obsidian, Mullumica source from Ecuador, five samples from Mullumica bottom field (CM1, CM2, CM3, CM4, and CM5), Fig. 19.16, and two samples from Mullumica upper field (CSM1 and CSM2) [39, 40, and 43], Fig. 19.17.

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

539

Fig. 19.16 Mullumica obsidian samples in Ecuador (CM1, CM2, CM3, CM4, and CM5) analyzed by portable X-ray fluorescence equipment, PXRF-LFNA-02 and PXRF-LFNA-03

Fig. 19.17 Mullumica obsidian samples in Ecuador (CSM1 and CSM2) analyzed by portable X-ray fluorescence equipment, PXRF-LFNA-02 and PXRFLFNA-03

Figure 19.18 shows a photo of the four reference (standard) obsidian samples used in this work for the construction of Ti, Fe, Sr, Rb, and Zr element quantification curves.

Portable System of EDXRF in Obsidian Analysis The PXRF-LFNA-02 system, used for analysis of elements with atomic number greater than 26, is composed of a 4 W X-ray tube (with Ag filter and target) and a Si-PIN detector model XR-100CR of Amptek Inc., which has a resolution of 221 eV

540

C. R. Appoloni et al.

Fig. 19.18 Certified obsidian samples analyzed by portable X-ray fluorescence equipment, PXRFLFNA-02 and PXRF-LFNA-03

for the 5.9 keV line (25 μm-thickness Be window and Ag collimator) [9]. The system used for the analysis of elements with atomic number lower than 26 was the PXRFLFNA-03, composed of a 4 W X-ray tube with W target and a Si-PIN detector, model XR-100CR of Amptek Inc., which has a resolution of 149 eV for the 5.9 keV line (12.7 μm-thickness Be window and Ag collimator) [9]. The measurement conditions were defined through the factorial design carried out, and acquisition time spectrum was 1000 s.

Optimization of pXRF Systems for Obsidian Analysis For obsidian analysis, the analytical sensitivity of mobile devices has been optimized using a factorial design 24 to determine elemental chemistry of archaeological objects in the laboratory and in situ, especially obsidian. For the optimization of the portable fluorescent X-ray energy dispersion (EDXRF) systems, the variables studied were (1) distance between sample and detector, (2) distance between sample and X-ray tube, (3) current applied at X-ray tube, and (4) voltage applied at X-ray tube. Table 19.7 shows the four factors resulting from the screening performed and their respective levels chosen for the PXRF-LFNA-03 system according to the experimental optimization tests. The experimental tests were performed intensively by setting one variable and by varying all others. This procedure was repeated for the four variables, and it was possible to assemble an extensive set of data, which were thoroughly analyzed for later choice of the levels presented in Table 19.7. The same procedure was performed for the levels presented in Table 19.8 referring to the equipment used which was PXRF-LFNA-02. After the established factors and their different levels chosen in experimental tests, a complete factorial design 24 of two levels was executed in order to facilitate the investigation of the four factors, since this statistical analysis must be performed in at least two levels, since it observes a possible variation in the response of interest as the factors are manipulated varying in level. The result of the factorial planning 24 showed that for the PXRF-LFNA-02 system to maximize the desired responses, factor 1 (detector position) should remain at its upper (+) level, factor 2 (X-ray tube) should act at its lower ()

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

Table 19.7 Significant factors and their respective levels for the factorial design with the equipment PXRF-LFNA-03

System PXRF-LFNA-03

Table 19.8 Significant factors and their respective levels for the factorial design with the equipment PXRF-LFNA-02

System PXRF-LFNA-02

Factors 1 detector position (x) 2 X-ray tube position (y) 3 electric tension (U) 4 electric current (i)

Factors 1 detector position (x) 2 X-ray tube position (y) 3 electric tension (U) 4 electric current (i)

541

Levels () 2.2 cm 1.9 cm 30 kV 4 μA

(+) 2.3 cm 2.0 cm 40 kV 5 μA

Levels () 1.5 cm 1.8 cm 28 kV 10 μA

(+) 1.8 cm 1.3 cm 35 kV 12 μA

level, and factors 3 (electrical voltage in the X-ray tube) and 4 (electrical current in the X-ray tube) should act at their upper (+) levels. For the PXRF-LFNA-03, factor 1 (detector position) should remain at its lowest level (), factor 2 (X-ray tube position) should act at its upper (+) level, and already the factors 3 (electrical voltage in the X-ray tube) and 4 (electric current in the X-ray tube) must act in their lower levels ().

19.6.2 Results and Discussions Pottery Fragments Analysis Based on the results obtained from these measures, it was possible to observe elements K, Ca, Ti, Mn, Fe, Zn, Br, Rb, Sr, Y, Zr, and Pb in different analyzed fragments. The statistic deviations were, in general, 1% for Fe; 5% for Ca; 5% to 10% for Sr, Zr, Mn, Ti, and Zn; and 20–25% for K, Br, Rb, Y, and Pb. The elements Ca, Ti, Mn, Fe, Zn, Sr, and Zr were observed in the ceramic paste of all 68 pottery fragments analyzed, indicating that these elements are present in the composition of the clay used in manufacturing of these ceramics. The elements Fe, Sr, Mn, Ti, and Zn presented systematically higher intensities in the faces in relation to the ceramic paste in 43 out of the 68 analyzed fragments, indicating that there is a treatment in the faces, with enriching for those elements, which is the engobe [34]. For the grouping analysis, the net area averages of the chemical elements obtained in the ceramic bulk of all analyzed fragments were used, since the ceramic bulk characterizes the source of the clay used for manufacturing the artifacts and allows for identification of possible groupings of the fragments that were

542

C. R. Appoloni et al.

manufactured with the same clay source. In order to build the PCA, the MATLAB ® 2009a [38] software was used. Figures 19.19 and 19.20 illustrate the PCA results of the fragments from the surface and 113 cm; 10 to 20 cm, 132 cm, and 144 cm levels obtained with pXRF analysis. In Fig. 19.19, the formation of two groups can be noted. These groups consist of fragments of surface and 113 cm. The differentiation of the two groups occurs due to the high Ca concentration of the fragments of the 113 cm and the low Zn concentration of the surface samples. From Fig. 19.20, separation of fragments of stratigraphies 10–20 cm, 132 cm, and 144 cm may be visualized. In the case, the differentiation occurs due to differences in Zn and Sr element concentrations among the three groups of fragments. The separation between these groups of samples with same stratigraphy indicates that each group of fragments was manufactured with different clays.

Obsidian Analysis All these elements were quantified in this work by the calibration curves, which were systematically evaluated. In general, the calibration curves were statistically significant, given the analysis of variance (ANOVA) for the ranges of concentrations set by the reference samples. Figure 19.21 shows the calibration line for the rubidium element (Rb) obtained through the measurements of the certified samples using the PXRF-LFNA-02 equipment. According to the data presented in Table 19.9, it can be verified that the regression of the model applied to the calibration curve of the rubidium element (Rb) is statistically significant, since AQR/AQr ¼ 371.02, which compared to the value of the F to the 99% confidence level, F1,18.99% ¼ 8.29, reveals that AQR/AQr >> F1,18.99%. This ensures that the regression performed is significant for the range of values considered

Fig. 19.19 Principal component analysis (PCA) with pottery fragments excavated in two stratigraphies, surface and 113 cm

19

Laboratory Portable X-Ray Fluorescence (pXRF) Systems Design and. . .

543

Fig. 19.20 Principal component analysis (PCA) with pottery fragments excavated in three stratigraphies, 10–20 cm, 132 cm, and 144 cm

Fig. 19.21 Calibration curve for element Rb (equipment PXRF-LFNA-02)

where the regression line is significant. Likewise, the model used for the calibration curve does not present a lack of adjustment, since AQLack/MQPe ¼ 0.37. Comparing with F2,16.99% ¼ 6.23, we find that AQLack/MQPe 3 cm) samples, the transmission images were found to be degraded, because the Terahertz signal was strongly attenuated during propagation in the sample. A frequency-modulated continuous Terahertz scanner (based on electronic generation and detection of the signal) was preferable for larger (and thicker) samples and when spectroscopic information was

602

I. Cacciari

Fig. 21.12 Terahertz image of a bird mummy in transmission geometry. (Reprinted from [19])

Fig. 21.13 Selected Terahertz waveforms transmitted from (a) torso and (b) leg regions of a bird mummy. (Reprinted from [19])

not crucial [13, 18]. This kind of setup was found to be efficient also in discriminating human tissues: in particular at 100 GHz, bone and tendon can be distinguished from the surrounding tissues. On the other hand, at higher frequencies

21

Terahertz Waves in Archaeology

603

(~300 GHz), the resolution could be increased at the expense of a decrease in the transmitted signal. The ability of Terahertz waves to penetrate into objects and thus to generate reflection imaging can offer fascinating information on the wrapping technique of Egyptian mummies [11, 21, 23]. Even if the depth probed is generally quite limited, the strong reflections observed very close to the surface make it possible to deduce some details of the bandage folding. Moreover, it is interesting to note that the signal reflected by the fabric layers in comparison with other multilayer glued sample (such as glued paper) was found to be extremely attenuated. This may suggest that Terahertz pulse can penetrate the outermost fabric layers and that it is absorbed by inner resin-fabric layer.

21.4

Conclusions and Future Perspectives

In the last decades, Terahertz technology has been growing at an incredible speed, as demonstrated by the increasing number of published papers and conference presentations. The researchers have been in pursuit of important technological advances since the first demonstration of Terahertz spectroscopy in the late 1980s. This challenge has provided new opportunities for understanding the basic science in this frequency band. A good knowledge of how to generate and detect Terahertz waves as well as the interaction with matter is crucial in order to use Terahertz in the best conditions. The positioning of Terahertz waves between infrared and microwaves domains allows to share some of their characteristics. In particular, since Terahertz waves have low energies, they are considered non-ionizing; this means that they are safe for biological tissues. This opens up a big opportunity for Terahertz imaging in medical applications. Moreover, Terahertz waves can penetrate through many dielectric materials, and as an immediate application, they can be used for non-destructive sensing and inspection, e.g., in industrial applications. On the other hand, the main limitations are that Terahertz waves cannot pass through metal and are strongly attenuated by water. However, the combination of all the features mentioned before provides the THz radiations with a great potential in many applications. This boosts the range of application beyond medicine and industry. It is further reinforced by the development of components, for generation and detection of Terahertz waves that are based on photonic technology. Although these strategic components are still expensive and bulky [154], the interest in developing new Terahertz applications has not been limited since they are nowadays commercially available. The great expectations that generally accompany Terahertz waves have determined the development of very interesting applications: these may include novel Terahertz integrated circuits for imaging applications [155], detection of cells and tissues [156], and wireless communications [157]. In the framework of innovative applications, it should be reported that Terahertz has been widely used for the study and characterization of different materials and

604

I. Cacciari

artworks. The success of Terahertz waves in the field of conservation, and more generally in that of cultural heritage, derives from the appealing possibility of providing information over extended areas by means of imaging and detailed data about the constituent chemicals of the artwork by means of spectroscopy. This is accomplished in a totally non-destructive perspective. Despite this success, the use of Terahertz waves in archaeology seems to be in its infancy. The few published works that join Terahertz waves with to the study of archaeological materials and artifacts are generally related to imaging of archaeological remains of different raw materials ranging from sealed pottery to metal artifacts and mummified tissues. Terahertz waves introduce a new way of “seeing” inside the object: in sealed objects, it was possible to obtain information on its content; in metal artifacts, it was possible to study the corrosion layers in order to locate the presence of the original metal; in mummified tissues, Terahertz waves were able to discriminate different tissues. In conclusion, we believe that the analysis should be extended over the plethora of different materials and artworks and should be encouraged. In this way, new answers to questions concerning materials characterization, which have been left unsolved by complementary techniques, can be offered to the archaeological community.

References 1. Sun Q, He Y, Liu K, Fan S, Parrott EPJ, Pickwell-MacPherson E (2017) Recent advances in terahertz technology for biomedical applications. Quant Imaging Med Surg 7(3):345–355 2. Nuss MC, Orenstein J (1997) Terahertz time domain spectroscopy (THz-TDS). In: Gruener G (ed) Millimeter-wave spectroscopy of solids. Springer, Berlin 3. Mittleman DM (ed) (2003) Sensing with Terahertz radiation. Springer, Berlin 4. Ferguson B, Zhang X-C (2002) Materials for terahertz science and technology. Nat Mater 1: 26–33 5. Dhillon SS, Vitiello MS, Linfield EH, Davies AG, Hoffmann MC, Booske J, Paoloni C, Gensch M, Weightman P, Williams GP, Castro-Camus E, Cumming DRS, Simoens F, Escorcia-Carranza I, Grant J, Lucyszyn S, Kuwata-Gonokami M, Konishi K, Koch M, Schmuttenmaer CA, Cocker TL, Huber R, Markelz AG, Taylor ZD, Wallace VP, Zeitler JA, Sibik J, Korter TM, Ellison B, Rea S, Goldsmith P, Cooper KB, Appleby R, Pardo D, Huggard PG, Krozer V, Shams H, Fice M, Renaud C, Seeds A, Stöhr A, Naftaly M, Ridler N, Clarke R, Cunningham JE, Johnston MB (2017) The 2017 terahertz science and technology roadmap. J Phys D Appl Phys 50(4):043001 6. Öhrström L, Bitzer A, Walther M, Rühli FJ (2010) Technical note: terahertz imaging of ancient mummies and bone. Am J Phys Anthropol 142(3):497–500 7. Jackson JB, Labaune J, Mourou G, Duling IN, Walker G, Bowen J, Menu M (2011) Terahertz pulse imaging of stratified architectural materials for cultural heritage studies. In: Proc. SPIE 2011. SPIE Optical Metrology, Munich, p 808409 8. Jackson JB, Labaune J, Mourou GA, D’Alessandro L, Whyte A, Menu M (2011) Pulsed terahertz investigation of corroded and mineralized copper alloy historical artifacts. In: Proceedings 2011 international conference on Infrared, Millimeter, and Terahertz waves. IrmmwTHz, Huston, pp 1–2 9. Jackson JB, Bowen J, Walker G, Labaune J, Mourou G, Menu M, Fukunaga K (2011) A survey of terahertz applications in cultural heritage conservation science. IEEE Trans Terahertz Sci Technol 1(1):220–231

21

Terahertz Waves in Archaeology

605

10. Caumes J-P, Younus A, Salort S, Chassagne B, Recur B, Ziéglé A, Dautant A, Abraham E (2011) Terahertz tomographic imaging of XVIIIth Dynasty Egyptian sealed pottery. Appl Opt 50(20):3604–3608 11. Fukunaga K, Cortes E, Cosentino A, Stünkel I, Leona M, Duling IN, Mininberg DT (2011) Investigating the use of terahertz pulsed time domain reflection imaging for the study of fabric layers of an Egyptian mummy. J Eur Opt Soc-Rapid 6:11040 12. J. Labaune, J.B. Jackson, K. Fukunaga,J. White, L. d’Alessandro, A. Whyte,M. Menu, G. Mourou: Investigation of Terra Cotta artefacts with terahertz, Appl Phys A Mater Sci Process 105, 5–9 (2011) 13. Younus A, Caumes J-P, Salort S, Chassagne B, Pradère C, Dautant A, Ziéglé A, Abraham E (2011) A continuous millimeter-wave imaging scanner for art conservation science. Adv At Mol Opt Phys 2011:275682 14. Bessou M, Duday H, Caumes J-P, Salort S, Chassagne B et al (2012) Advantage of terahertz radiation versus X-ray to detect hidden organic materials in sealed vessels. Opt Commun 285(21–22):4175–4179 15. Walker GC, Bowen JW, Matthews W, Roychowdhury S, Labaune J, Mourou G, Menu M, Hodder I, Jackson JB (2013) Sub-surface terahertz imaging through uneven surfaces: visualizing Neolithic wall paintings in Çatalhöyük. Opt Express 21(7):8126–8134 16. Abraham E, Bessou M, Ziéglé A, Hervé M-C, Szentmiklósi L, Kasztovszky ZS, Kis Z, Menu M (2014) Terahertz, X-ray and neutron computed tomography of an eighteenth Dynasty Egyptian sealed pottery. Appl Phys A Mater Sci Process 117(3):963–972 17. Sasaki T, Hashimoto Y, Mori T, Kojima S (2015) Broadband terahertz time-domain spectroscopy of archaeological Baltic Amber. Int Lett Chem Phys Astronomy 62:29–33 18. Öhrström L, Fischer BM, Bitzer A, Wallauer J, Walther M, Rühli F (2015) Terahertz imaging modalities of ancient Egyptian mummified objects and of a naturally mummified rat. Anat Rec (Hoboken) 298(6):1135–1143 19. Jackson JB, Labaune J, Bailleul-Lesuer R, D'Alessandro L, Whyte A, Bowen JW, Menu M, Mourou G (2015) Terahertz pulse imaging in archaeology. Front Optoelectron 8(1):81–92 20. Cacciari I, Agresti J, Siano S (2016) Combined THz and LIPS analysis of corroded archaeological bronzes. Microchem J 126:76–82 21. Jackson B, Mourou G, Labaune J, Menu M (2015) Terahertz pulse imaging of an Egyptian Bird mummy. In: Bailleul-LeSuer R (ed) Between heaven and earth, Birds in Ancient Egypt. Oriental Institute Museum Publications, Chicago, pp 119–122 22. Jackson JB (2016) TISCH – Terahertz imaging and spectroscopy in cultural heritage: applications in archeology, architecture and art conservation science. Proceeding optical sensors. Optical Society of America, Vancouver, pp SeTu3E.3(1)–(3)) 23. Fukunaga K, Kohdzuma Y (2016) THz pulsed time domain imaging applied to museum objects and its comparison with other frequency regions. Proceeding 19th world conference on non-destructive testing. Curran Associates, Inc., New York, pp 676–683 24. Cacciari I, Siano S (2017) Use of THz reflectometry for roughness estimations of archaeological metal surfaces. J Infrared Milli Terahz Waves 38:503–517 25. Faustino MAB, Lopez Jr LP, Cabello NI, De Los Reyes A, Cafe A, Tugado C, Hernandez V, Barretto-Tesoro G, Estacio ES (2017) Terahertz transmission of soil samples excavated from Pinagbayanan, San Juan, Batangas. Proceedings of the Samahang Pisika ng, Pilipinas. Cebu City, Philippine, pp SPP-2017-2D-03-1(4)) 26. Amir F, Mitchel CJ, Farrington N, Missous M (2009) Advanced Gunn diode as high power terahertz source for a millimetre wave high power multiplier. In: Proc SPIE 2009. SPIE Security + Defence, Berlin, p 74850I(1)–(11) 27. Pérez S, González T, Pardo D, Mateos J (2008) Terahertz Gunn-like oscillations in InGaAs/ InAlAs planar diodes. J Appl Phys 103:0945160 28. Mukherjee M, Mazumder N, Roy SK, Goswami K (2007) GaN IMPATT diode: a photosensitive high power terahertz source. Semicond Sci Technol 22(12):1258–1267 29. Acharyya A, Banerjee JP (2014) Prospects of IMPATT devices based on wide bandgap semiconductors as potential terahertz sources. Appl Nanosci 4(1):1–14

606

I. Cacciari

30. E. Wasige, K.H. Alharbi, A. Al-Khalidi, J. Wang, A. Khalid, G.C. Rodrigues, J. Figueiredo: Resonant tunnelling diode terahertz sources for broadband wireless communications, Proceedings Volume 10103, Terahertz, RF, Millimeter, and Submillimeter-Wave Technology and Applications X, (SPIE OPTO, San Francisco 2017) pp. 101031J(1)–(11) 31. Asada M, Suzuki S, Kishimoto N (2008) Resonant tunneling diodes for sub-terahertz and terahertz oscillators. Jpn J Appl Phys 47(6):4375–4384 32. Orihashi N, Suzuki S (2005) One THz harmonic oscillation of resonant tunneling diodes. Appl Phys Lett 87:233501 33. Udal A, Jaanus M, Valušis G, Kašalynas I, Ikonic Z, Indjin D (2015) Progress in development of the resonant tunneling diodes as promising compact sources at the THz gap bottom. In: Pereira MF, Shulika O (eds) THz for CBRN and explosives detection and diagnosis. NATO science for peace and security series B: physics and biophysics. Springer, Dordrecht, pp 169– 178 34. Byrd JM, Leemans WP, Loftsdottir A, Marcelis B, Martin MC, McKinney WR, Sannibale F, Scarvie T, Steier C (2002) Observation of broadband self-amplified spontaneous coherent terahertz synchrotron radiation in a storage ring. Phys Rev Lett 89:224801 35. Mineo M, Paoloni C (2010) Corrugated rectangular waveguide tunable backward wave oscillator for terahertz applications. IEEE Trans Electron Devices 57(6):1481–1484 36. Xu X, Wei Y, Shen F, Yin H, Xu J, Gong Y, Wang W (2012) A watt-class 1-THz backwardwave oscillator based on sine waveguide. Phys Plasmas 19:013113(1)–(5) 37. He W, Donaldson CR, Zhang L, Ronald K, McElhinney P, Cross AW (2013) High power wideband gyrotron backward wave oscillator operating towards the terahertz region. Phys Rev Lett 110(16):165101(1)–(5) 38. Gallerano GP, Biedron S. Overview of terahertz radiation sources. Proceedings of the 26th international free electron laser conference and 11th FEL Users workshop, pp 216–221 39. Köhler R, Tredicucci A, Beltram F, Beere HE, Linfield EH, Davies AG, Ritchie DA, Iotti RC, Rossi F (2002) Terahertz semiconductor-heterostructure laser. Nature 417:156–159 40. Hübers H-W, Pavlov SG, Shastin VN (2005) Terahertz lasers based on germanium and silicon Semicond. Sci Technol 20:S211–S221 41. Odnoblyudov MA, Prokofie AA, Yassievich IN, Chao KA (2004) Theory of a strained p-Ge resonant-state terahertz laser. Phys Rev B 70:115209(1)–(14) 42. Chassagneux Y, Colombelli R, Maineult W, Barbieri S, Beere HE, Ritchie DA, Khanna SP, Linfield EH, Davies AG (2009) Electrically pumped photonic-crystal terahertz lasers controlled by boundary conditions. Nature 457:174–178 43. Brenner C, Hoffmann S, Friedrich C-S, Schlauch T, Klehr A, Erbert G, Tränkle G, Jördens C, Salhi M, Koch M, Hofmann MR (2009) Semiconductor laser based THz generation and detection. Phys Status Solidi C 6(2):564–567 44. Vijayraghavan K, Jiang Y, Jung S, Belkin M (2004) Sensing & measurement, compact and broadly tunable semiconductor terahertz radiation sources. SPIE newsroom, pp 1–3 45. Brenner C, Hoffmann S, Hofmann M, Salhi M, Koch M (2006) Room-temperature terahertz detection with diode lasers. In: Conference on lasers and electro-optics. IEEE, Piscataway, p CTuL6(1)–(2) 46. Lu QY, Slivken S, Bandyopadhyay N, Bai Y, Razeghi M (2014) Widely tunable room temperature semiconductor terahertz source. Appl Phys Lett 105(20):201102(1)–(5) 47. Faist J, Capasso F, Sivco DL, Sirtori C, Hutchinson AL, Cho AY (1994) Quantum Cascade Laser Sci 264(5158):553–556 48. Williams BS, Kumar S, Hu Q, Reno JL (2005) Operation of terahertz quantum-cascade lasers at 164K in pulsed mode and at 117K in continuous-wave mode. Opt Express 13(9):3331–3339 49. Williams BS (2007) Terahertz quantum-cascade lasers. Nature Photon 1:517–525 50. S. Kumar, Q. Hu, and J.L. Reno: 186K operation of terahertz quantum-cascade lasers based on a diagonal design, Appl Phys Lett. 94 131105(1)–(3) (2009) 51. Scalari G, Walther C, Fischer M, Terazzi R, Beere H, Ritchie D, Faist J (2009) THz and sub-THz quantum cascade lasers. Laser Photonics Rev 3(1–2):45–66

21

Terahertz Waves in Archaeology

607

52. Mechold L, Kunsh J (2004) QCL modules are ready for industrial application. Laser Focus World 40(5):88–92 53. AzoOptics. https://www.azooptics.com/Article.aspx?ArticleID¼1403 54. Maldonado D. https://www.photonics.com/Articles/QCL_Primer_History_Characteristics_ Applications/a58343 55. Schwaighofer A, Brandstetter M, Lendl B (2017) Quantum cascade lasers (QCLs) in biomedical spectroscopy. Chem Soc Rev 46:5903–5924 56. Sreedhar H, Varma VK, Bird B, Guzman G, Walsh MJ (2017) Applications of QCL mid-IR imaging to the advancement of pathology. Proceedings of the Volume 10140, medical imaging 2017: digital pathology. SPIE Medical Imaging, Orlando, pp 1014004(1)–(7) 57. Dean P, Valavanis A, Keeley J, Bertling K, Lim YL, Alhathlool R, Burnett AD, Li LH, Khanna SP, Indjin D, Taimre T, Rakić AD, Linfield EH, Davies AG (2014) Terahertz imaging using quantum cascade lasers-a review of systems and applications. J Phys D Appl Phys 47:374008(1)–(22) 58. Darmo J, Tamosiunas V, Fasching G, Kröll J, Unterrainer K, Beck M, Giovannini M, Faist J, Kremser C, Debbage P (2004) Imaging with a terahertz quantum cascade laser. Opt Express 12(9):1879–1884 59. Nguyen KL, Johns ML, Gladden LF, Worrall CH, Alexander P, Beere HE, Pepper M, Ritchie DA, Alton J, Barbieri S, Linfield EH (2006) Three-dimensional imaging with a terahertz quantum cascade laser. Opt Express 14(6):2123–2129 60. Fukunaga K (2016) THz technology applied to cultural heritage in practice. Springer, Tokyo 61. Auston DH, Cheung KP, Valdmanis JA, Kleinman DA (1984) Cherenkov radiation from femtosecond optical pulses in electro-optic media. Phys Rev Lett 53(16):1555–1558 62. Fattinger C, Grischkowsky D (1989) Terahertz beams. Appl Phys Lett 54(6):490–492 63. Jiang Z, Zhang X-C (1999) Terahertz imaging via electrooptic effect. IEEE Trans Microwave Theory Tech 47(12):2644–2650 64. Wu Q, Hewitt TD, Zhang X-C (1996) Two-dimensional electro-optic imaging of THz beams. Appl Phys Lett 69(8):1026–1028 65. Brown ER, McIntosh KA, Nichols KB, Dennis CL (1995) Photomixing up to 3.8 THz in low temperature-grown GaAs. Appl Phys Lett 66(3):285–287 66. McIntosh KA, Brown ER, Nichols KB, McMahon OB, DiNatale WF, Lyszczarz TM (1995) Terahertz photomixing with diode lasers in low-temperature-grown GaAs. Appl Phys Lett 67(26):3844–3846 67. Yin X, Ng BW-H, Abbott D (2012) Terahertz sources and detectors. In: Yin X, Ng BW-H, Abbott D (eds) Terahertz imaging for biomedical applications: pattern recognition and tomographic reconstruction. Springer, New York 68. Nellen S, Globisch B, Kohlhaas RB, Liebermeister L, Schell M (2018) Recent progress of continuous-wave terahertz systems for spectroscopy, non-destructive testing, and telecommunication. Proceedings of the SPIE 10531, Terahertz, RF, millimeter, and submillimeter-wave technology and applications XI, C. SPIE, San Francisco, pp 105310(1)–(9) 69. Hindle F, Cuisset A, Bocquet R, Mouret G (2008) Continuous-wave terahertz by photomixing: applications to gas phase pollutant detection and quantification. C R Physique 9(2):262–275 70. Park I, Sydlo C, Fischer I, Elsäßer W, Hartnage HL (2008) Generation and spectroscopic application of tunable continuous-wave terahertz radiation using a dual-mode semiconductor laser. Meas Sci Technol 19:065305(1)–(9) 71. Dong SW, Zhu Z-B, Wang Y (2009) CW terahertz source techniques for space systems. Proceedings Volume 7385, International Symposium on Photoelectronic Detection and Imaging 2009. SPIE, Beijing, pp 73851A(1)–(6) 72. Brown ER, Segovia-Vargas D (2015) Principles of THz direct detection. In: Carpintero G, Muñoz LEG, Hartnagel HL, Preu S, Räisänen AV (eds) Semiconductor Terahertz technology: devices and systems at room temperature operation. Wiley, Hoboken 73. Lee AWM, Qin Q, Kumar S, Williams BS, Hu Q, Reno JL (2006) Appl Phys Lett 89(14): 141125(1)–(3)

608

I. Cacciari

74. Lee AWM, Williams BS, Kumar S, Hu Q, Reno JL (2006) Real-time imaging using a 4.3-THz quantum cascade laser and a 320 240 microbolometer focal-plane array. IEEE Photonics Technol Lett 18(13):1415–1417 75. Hammar A, Cherednichenko S, Bevilacqua S, Stake J (2011) Terahertz direct detection in microbolometers. IEEE Trans Terahertz Sci Technol 1(2):390–394 76. Dobroiu A, Yamashita M, Ohshima YN, Morita Y, Otani C, Kawase K (2004) Terahertz imaging system based on a backward-wave oscillator. Appl Opt 43(60):5637–5646 77. Lee Y-S (2009) Principles of terahertz science and technology. Springer, New York 78. Hargreaves S, Lewis RA (2007) Terahertz imaging: materials and methods. J Mater Sci Mater Electron 18:S299–S303 79. Karpowicz N, Zhong H, Xu J, Lin K-I, Hwang J-S, Zhang X-C (2005) Nondestructive sub-THz imaging, proceedings volume 5727, terahertz and gigahertz electronics and photonics IV. SPIE, San Jose, pp 132–142 80. Sizova F, Rogalski A (2010) THz detectors. Prog Quant Electron 34(5):278–347 81. Roser HP, Hubers HW, Crowe TW, Peatman WCB (1994) Nanostructure GaAs Schottky diodes for far-infrared heterodyne receivers. IR Physics 35(2–3):451–462 82. Siegel PH, Mehdi I (1996) The spare-diode-detector: a new planar diode detector circuit with state-of-the-art performance for millimeter and submillimeter wavelengths. JPL New Technol Rep NPO-20 104 83. Erickson N (2002) A fast, very sensitive calorimetric power meter for millimeter to submillimeter wavelengths. Proceedings of the 13th international space terahertz technology symp, pp 301–308 84. Turner AD, Bock JJ, Beeman JW, Glenn J, Hargrave PC, Hristov VV, Nguyen HT, Rahman F, Sethuraman S, Woodcraft AL (2001) Silicon nitride micromesh bolometer array for submillimeter astrophysics. Appl Opt 40(28):4921–4932 85. Voellmer GM, Allen CA, Amato MJ, Babu SR, Bartels AE, Benford DJ, Derro RJ, Dowell CD, Harper DA, Jhabvala MD, Moseley SH, Rennick T, Shirron PJ, Smith WW, Staguhn JG (2003) Design and fabrication of twodimensional semiconducting bolometer arrays for HAWC and SHARCII. Proceedings of SPIE 4855, millimeter and submillimeter detectors for astronomy, pp 63–72 86. Staguhn JG, Benford DJ, Pajot F, Ames TJ, Chervenak JA, Grossman EN, Irwin KD, Maffei B, Moseley SH, Phillips TG, Reintsema CD, Rioux C, Shafer RA, Voellmer GM (2003) Astronomical demonstration of superconducting bolometer arrays. Proceedings SPIE 4855, millimeter and submillimeter detectors for astronomy, pp 100–107 87. Kim DY, Park S, Han R, Kenneth KO (2016) Design and demonstration of 820-GHz array using diode-connected NMOS transistors in 130-nm CMOS for active imaging. IEEE THz Sci Technol 6(2):306–317 88. Krishna MG, Kshirsagar SD, Tewari SP (2012) Terahertz emitters, detectors and sensors: current status and future prospects. IntechOpen, London, pp 115–144 89. Cai Y, Brener I, Lopata J, Wynn J, Pfeiffer L, Stark JB, Wu Q, Zhang XC, Federici JF (1998) Direct comparison between free-space electro-optic sampling and antenna detection. Appl Phys Lett 73(4):444–446 90. Ibrahim A, Férachou D, Sharma G, Singh K, Kirouac-Turmel M, Ozaki T (2016) Ultra-high dynamic range electro-optic sampling for detecting millimeter and sub-millimeter radiation. Scient Rep 6:23107(1)–(9) 91. Nahata A, Auston DH, Heinz TF, Wu C (1996) Coherent detection of freely propagating terahertz radiation by electro-optic sampling. Appl Phys Lett 68(2):150–152 92. Hu BB, Nuss MC (1995) Imaging with terahertz waves. Opt Lett 20(16):1716–1718 93. Jin Y-S, Jeon S-G, Kim G-J, Kim J-I, Shon C-H (2007) Fast scanning of a pulsed terahertz signal using an oscillating optical delay line. Rev Sci Instrum 78(2):023101(1)–(5) 94. Kim GY, Jeon S-G, Kim J-I, Jin Y-S (2008) High speed scanning of terahertz pulse by a rotary optical delay. Rev Scient Instr 79(10):106102(1)–(3)

21

Terahertz Waves in Archaeology

609

95. Ahi K, Anwar M (2016) Developing terahertz imaging equation and enhancement of the resolution of terahertz images using deconvolution. Proceedings SPIE, Terahertz physics, devices, and systems X: advanced applications in industry and defense, pp 98560N(1)– (18) 96. Cacciari I, Mugnai D, Ranfagni A (2019) Resolving power beyond the diffraction limit demonstrated with composed pupils at microwave and THz frequencies. J Appl Phys 125: 044901(1)–(7) 97. Lo YH, Leonhardt R (2008) Aspheric lenses for terahertz imaging. Opt Express 16:15991– 15998 98. Bandyopadhyay A, Stepanov A, Schulkin B, Federici MD, Sengupta A, Gary D, Federici JF, Barat R, Michalopoulou ZH, Zimdars D (2006) Terahertz interferometric and synthetic aperture imaging. JOSA A 23(5):1168–1178 99. Li X, Hong Z, He J, Chen Y (2010) Precisely optical material parameter determination by time domain waveform rebuilding with THz time-domain spectroscopy. Opt Commun 283(23): 4701–4706 100. Hua YF, Zhang HJ (2010) Qualitative and quantitative detection of pesticides with terahertz time-domain spectroscopy. IEEE Trans Microw Theory Tech 58(7):2064–2070 101. Gowen AA, O’Sullivan C, O’Donnell CP (2012) Terahertz time domain spectroscopy and imaging: emerging techniques for food process monitoring and quality control. Trends Food Sci Technol 25(1):40–46 102. Duvillaret L, Garet F, Coutaz JL (1996) A reliable method for extraction of material parameters in terahertz time-domain spectroscopy. IEEE J Sel Topics Quantum Electron 2(3):739–746 103. Duvillaret L, Garet F, Coutaz JL (1999) Highly precise determination of optical constants and samples thickness in terahertz time-domain spectroscopy. Appl Opt 38(2):409–415 104. Timothy DD, Richard GB, Daniel MM (2001) Material parameter estimation with terahertz time-domain spectroscopy. J Opt Soc Amer A Opt Image Sci 18(7):1562–1571 105. Mittleman DM, Jacobsen RH, Nuss MC (1996) T-ray imaging. IEEE J Sel Top Quantrum Electron 2(3):679–692 106. Brucherseifer M, Haring Bolivar P, Klingenberg H, Kurz H (2001) Angle dependent THz tomography characterization of thin ceramic oxide films for fuel cell applications. Appl Phys B Lasers Opt 72(2):361–366 107. Dobroiu A, Otani C, Kawase K (2006) Terahertz-wave sources and imaging applications. Meas Sci Technol 17(11):R161–R174 108. Yu C, Fan S, Sun Y, Pickwell-Macpherson E (2012) The potential of terahertz imaging for cancer diagnosis: a review of investigations to date, quant. Imaging Med Surg 2(1):33–44 109. Garbacz P (2016) Terahertz imaging – principles, techniques, benefits, and limitations. Problemy Eksploatacji 1:81–92 110. Ghann W, Uddin J. Terahertz (THz) spectroscopy: a cutting-edge technology. https://doi.org/ 10.5772/67031 111. Yamagiwa M, Ogawa T, Minamikawa T, Ghareab Abdelsalam D, Okabe K, Tsurumachi N, Mizutani Y, Iwata T, Yamamoto H, Yasui T (2018) Real-time amplitude and phase imaging of optically opaque objects by combining full-field off-axis terahertz digital holography with angular spectrum reconstruction. J Infrared Millim Terahertz Waves 39(6):561–572 112. Semenova VA, Kulya MS, Petrov NV, Grachev YV, Tsypkin AN, Putilin SE, Bespalov VG. Amplitude-phase imaging of pulsed broadband terahertz vortex beams generated by spiral phase plate. 2016 41st International conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), pp 1–2 113. Hack E, Zolliker P (2014) Terahertz holography for imaging amplitude and phase objects. Opt Express 22(13):16079–16086 114. Petrov NV, Kulya MS, Tsypkin AN, Bespalov VG, Gorodetsky A (2016) Application of terahertz pulse time-domain holography for phase imaging. IEEE T THZ Sci Technol 6(3):464–472

610

I. Cacciari

115. Bowman T, El-Shenawee M, Campbell LK (2016) Terahertz transmission vs reflection imaging and model-based characterization for excised breast carcinomas. Biomed Opt Express 7(9):3756–3783 116. Fan S, Parrott EPJ, Ung BSY, Pickwell-MacPherson E (2016) Calibration method to improve the accuracy of THz imaging and spectroscopy in reflection geometry. Photonic Res 4(3):A29– A35 117. Huang S, Ashworth PC, Chi W, Kan KWC, Chen Y, Wallace VP, Zhang Y-t, PickwellMacPherson E (2009) Improved sample characterization in terahertz reflection imaging and spectroscopy. Opt Express 17(5):3848–3854 118. Vandrevala F, Karmakar A, Jornet JM, Einarsson E (2016) Extracting complex optical properties of ultra-thin conductors using time-DOMAIN THz spectroscopy. 2016 41st International conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), pp 1–2 119. Bernier M, Garet F, Couta J-L (2016) Determining the complex refractive index of materials, in the far-infrared from terahertz time-domain data. InTech, London, pp 119–141 120. Vartiainen EM (2004) Numerical phase correction method for terahertz time-domain reflection spectroscopy. J Appl Phys 96(8):4171–4175 121. Naftaly M (2015) Evaluation of uncertainty in time-domain spectroscopy. In: Naftaly M (ed) Terahertz metrology. Artech House, Boston, pp 103–105 122. Pashkin A, Kempa M, Němec H, Kadlec F, Kužel P (2003) Phase-sensitive time domain terahertz reflection spectroscopy. Rev Sci Instrum 74(11):4711–4747 123. Jepsen PU, Møller U, Merbold H (2007) Investigation of aqueous alcohol and sugar solutions with reflection terahertz time domain spectroscopy. Opt Express 15(22):14717–14737 124. Redo-Sanchez A, Zhang XC (2011) Self-referenced method for terahertz wave time-domain spectroscopy. Opt Lett 36(17):3308–3310 125. Kawase K, Shibuya T, Suizu K, Hayashi S (2010) Terahertz-wave generation for industrial applications. Sensing & Measurement. http://spie.org/newsroom/2611-terahertz-wave-genera tion-for-industrial-applications?SSO¼1 126. Kawase K, Shibuya T, Hayashi S, Suizu K (2010) THz imaging techniques for nondestructive inspections. C R Physique 11(7–8):510–518 127. Ellrich F, Klier J, Weber S, Jonuscheit J, von Freymann G (2016) Terahertz time-domain technology for thickness determination of industrial relevant multi-layer coatings. 2016 41st International Conference on Infrared, Millimeter, and Terahertz waves (IRMMW-THz), pp 1–2 128. Zeitler JA, Shen YC, Baker C, Taday PF, Pepper M, Rades T (2007) Analysis of coating structures and interfaces in solid oral dosage forms by three dimensional terahertz pulsed imaging. J Pharm Sci 96(2):330–340 129. Schecklman S, Zurk LM (2018) Terahertz imaging of thin film layers with matched field processing. Sensors 18:1–20 130. Hauck J, Stich D, Heidemeyer P, Bastian M, Hochrein T (2014) Terahertz inline wall thickness monitoring system for plastic pipe extrusion. AIP Conference Proc 1593:86–89 131. Skryl AS, Jackson JB, Bakunov MI, Menu M, Mourou GA (2014) Terahertz time-domain imaging of hidden defects in wooden artworks: application to a Russian icon painting. Appl Opt 53(6):1033–1038 132. Krügener K, Busch SF, Soltani A, Castro-Camus E, Koch M, Viöl W (2017) Non-destructive analysis of material detachments from polychromatically glazed terracotta artwork by THz time-of-flight spectroscopy. J Infrared Millim Terahertz Waves 38(4):495–502 133. Arnone DD, Ciesla CM, Corchia A, Egusa S, Pepper M, Chamberlain JM, Bezant C, Linfield EH, Clothier R, Khammo N (1999) Applications of terahertz (THz) technology to medical imaging. Proceedings Volume 3828, Terahertz spectroscopy and applications II. SPIE, Munich, pp 209–219 134. Guillet JP, Recur B, Frederique L, Bousquet B, Canioni L, Manek-Hönninger I, Desbarats P, Mounaix P (2014) Review of Terahertz Tomography techniques. J Infrared Milli Terahz Waves 35(4):382–411

21

Terahertz Waves in Archaeology

611

135. Chan WL, Deibel J, Mittleman DM (2007) Imaging with terahertz radiation. Rep Prog Phys 70:1325–1379 136. Keith R (2018). https://www.quora.com/What-is-the-definition-of-artifacts-ecofactchronofacts-and-structures-in-archaeology 137. Peregrine PN (2017) The archaeological record. In: Peregrine PN (ed) Archaeological research: a brief introduction. New York, Routledge, p 15 138. Labaune J, Jackson JB, Pagès-Camagna S, Duling IN, Menu M, Mourou GA (2010) Papyrus imaging with terahertz time domain spectroscopy. Appl Phys A Mater Sci Process 100(3): 607–612 139. Uda M, Menortier G, Nakai I (eds) (2005) X-rays for archaeology. Springer, Berlin 140. Andreani C, Aliotta F, Arcidiacono L, Borla M, Di Martino D, Facchetti F, Ferraris E, Festa G, Gorini G, Kockelmann W, Kelleher J, Malfitana D, Micieli D, Minniti T, Perelli Cippo E, Ponterio R, Salvato G, Senesi R, Turina V, Vasi C, Greco C (2017) A neutron study of sealed pottery from the grave goods of Kha and MeritJ. Anal At Spectrom 32(7):1342–1347 141. McKenzie-Clark J, Magnussen J (2016) Real and virtual: the role of computed tomography and 3D imaging in museum practice. In: A cultural cacophony: museum perspectives and projects. NSW branch of Museums Galleries Australia, Sydney, pp 208–221 142. Beckman P, Spizzichino A (1987) The scattering of electromagnetic waves from rough surfaces. Artech House, London 143. Cacciari I, Ciofini D, Mascalchi M, Mencaglia A, Siano S (2012) Novel approach to the microscopic inspectrion during laser cleaning treatments of artworks. Anal Bional Chem 402(4):1585–1591 144. MantlerM M (2001) Schreiner: X-ray analysis of objects of art and archaeology. J Radioanal Nucl Chem 247(3):635–644 145. Schreiner M, Frühmann B, Jembrih-Simbürger D, Linke R (2004) X-rays in art and archaeology: an overview. Powder Diffract 19(1):3–11 146. Renfrew C, Shennan S (eds) (1982) Ranking, resource and exchange: aspect of the archaeology of Early European Society. Cambridge University Press, Cambridge 147. Anderson KB, Crelling JC (eds) (1995) Amber, resinite and fossil resins. American Chemical Society, Washington, DC 148. Shibata T, Mori T, Kjina S (2015) Low-frequency vibrational properties of crystalline and glassy indomethacin probed by terahertz time-domain spectroscopy and low-frequency Raman scattering, Spectrochim. Acta A 150:207–2011 149. Friesen TM, Maso OK (eds) (2016) The Oxford handbook of the prehistoric Arctic. Oxford University Press, New York 150. Henderson J (ed) (2000) The science and archaeology of materials: an investigation of inorganic materials. Routledge, New York 151. De Boer HH, Van Der Merwe AE (2016) Diagnostic dry bone histology in human paleopathology. Clin Anat 29(7):831–843 152. Fernández PL (2012) Palaeopathology: the study of disease in the past. Pathobiology 79(5): 221–227 153. Metcalfe NH (2007) In what ways can human skeletal remains be used to understand health and disease from the past? Postgrad Med J 83(978):281–284 154. Hochrein T (2015) Markets, availability, notice, and technical performance of terahertz systems: historic development, present, and trends. J Infrared Milli Terahz Waves 36(3): 235–254 155. Hillger P, Grzyb J, Jain R, Pfeiffer UR (2019) Terahertz imaging and sensing applications with silicon-based technologies. IEEE Trans Terahertz Sci Technol 9(1):1–19 156. Yu L, Hao L, Meiqiong T, Jiaoqi H, Wei L, Jinying D, Xueping C, Weiling F, Yang Z (2019) The medical application of terahertz technology in non-invasive detection of cells and tissues: opportunities and challenges. RSC Adv 9(17):9354–9363 157. Akkaş MA (2019) Terahertz wireless data communication. Wirel Netw 25(1):145–155

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

22

Anabelle Kriznar

Contents 22.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.1.1 Techniques of Mediaeval Mural Paintings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.1.2 Pigments and Binders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.1.3 Painting Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.1 Precise Examination by the Naked Eye with the Help of Raking Light . . . . . . 22.3.2 Digital Microscopes and Endoscopes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.3 Ultraviolet and Infrared Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.4 Portable X-Ray Fluorescence (XRF) Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3.5 Portable VIS Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

614 616 618 619 622 622 623 627 631 633 636 637 638

Abstract

In many mediaeval churches through Europe, mural paintings have been (and still are) discovered. Due to different atmospheric and building factors, they can be in a bad conservation state. For their monitoring and conservation, it is necessary to know materials and painting techniques used by mediaeval artists/workshops. In order to get as complete information as possible, the best way is interdisciplinary research between humanists and natural scientists, to combine the knowledge about the style, history, current conservation situation, materials, and their possible later changes. There are several approaches possible which are completely noninvasive, from a precise examination by the naked eye, using different light sources and small digital microscopes, UV and IRR radiation, to material A. Kriznar (*) Departamento de Escultura e Historia de las Artes Plásticas, Facultad de Bellas Artes, Centro Nacional de Aceleradores, Universidad de Sevilla, Seville, Spain Department of Art History, Faculty of Arts, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_22

613

614

A. Kriznar

characterization. The most used method is X-ray fluorescence (XRF) that identifies chemical elements in analyzed points, and therefore identifies pigments and supports in the mural painting. It can be, sometimes, combined with XRD. A helpful technique is also VIS spectrophotometry, which identifies the color of a paint layer, in case they are doubts – due to darkness, dirt on the surface, or pigment change. Keywords

Mural painting · 14th century · Plasters · Pigments · Painting techniques · Painting procedures · Noninvasive analysis · Digital analysis · UV-IRR · XRF

22.1

Introduction

Mural or wall paintings are artworks carried out on a wall (inner or outer) of a building; therefore they are considered immobile cultural heritage. They can be, of course, detached and conserved in a controlled environment as a museum, but mostly this is not the case – they continue forming part of the architecture through centuries. This means, that they share environmental conditions inside and outside the building, and especially if they are situated on outside walls, they suffer from strong weather changes (rain, snow, ice, sun, high temperatures). Humidity can cause many damages, as well as also bacterial and fungus attack, which can destroy color layers and their support. We should not forget about the human factor – vandalism, war destruction, and sometimes, as well, bad restoration interventions. In mediaeval times, wall paintings mostly decorated church walls, acting as the Bible in images for people that could not read (most of the population outside bigger towns). Sometimes also castles, palaces, and other rich private houses were decorated with wall paintings, but nobles have generally chosen profane stories from court life, to make the interiors more comfortable, showing off also their social and economic power. In any of those buildings, sacral or profane, paintings often suffered many mishandlings through the centuries – due to taste/style changes. Especially in the Baroque, most of mediaeval paintings were covered by lime wash or even hidden under another plaster layer; they could get hammered, so that the new plaster could better bond to the older one. Many reconstructions of buildings were carried out, turning mediaeval churches and palaces into Renaissance or Baroque ones, during which also paintings were totally or partially destroyed. Only from the nineteenth century, “old things” got their value back and many hidden paintings were (and still are) brought to light, removing lime wash, plaster, and sometimes also younger paintings. Nevertheless, also during the removal, damages can happen – color layers can be too strongly bind to the lime wash or plaster, so with their removal, also the paintings can be lost (Fig. 22.1). Through the centuries and due to the changing atmospheric conditions mentioned above, murals suffer not only from the humidity in the air; capillary ascending moisture drawn from the ground is also very dangerous for the plasters and painting

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

615

Fig. 22.1 Damaged mural paintings due to a later overlaid plaster, removed during the restoration. Hammering can still be observed, and the lower part of the painting is lost. Plešivec, Slovakia (1370– 1380)

layers. The porous walls suck up moisture from the earth, due to the pressure which is developed because of the surface tension of the moisture and the capillaries/pores in the building material, and this transports the moisture upward. Even bigger problem is that this humidity also transports harmful salts from the earth, which causes the plaster to flake off and nitrate efflorescence to appear. The more delicate the capillary, the more pressure is generated and the higher the water with salts mounts. This originates decomposition of the basic building structure, the plaster, and with this, the paintings [1, 2]. The new problem of the twentieth and the twentyfirst centuries is also the air pollution, which do not impact only live beings but also monuments. The presence of hydrogen sulfide (H2S) or sulfur dioxide (SO2) atmospheric or biological origin can cause pigment deterioration and its darkening, which could be also, on the other hand, a result of too high temperature (fire; candles that were lit too close to the painting) [3]. All these factors have a huge influence on state of the conservation of the murals, but this depends also on other two important factors: materials (lime, sand for plaster, pigments, binders for colors) and painting techniques applied by the artists and their workshops. Not all materials are suitable for every painting technique and for every environment; some are not very stable and can degrade quickly, especially organic colorants and binders. Knowing the materials and techniques used by old masters, understanding later damages and

616

A. Kriznar

weathering conditions, the conservators and restorers can do much better work protecting of our cultural heritage and preserving it for future generations.

22.1.1 Techniques of Mediaeval Mural Paintings Mural paintings can be carried out on a fresh plaster, on a fresh lime wash, or on already dried plaster; therefore there are basically three different painting techniques: a fresco, lime technique, and a secco. Painting on a fresh mortar, fresco buono, was for sure already known in Ancient Rome; however it is possible that it was used even before, in Greek painting or even in the Aegean culture around 15th BC. After the fall of the Roman Empire, it was forgotten in Europe but was continuously used by the artist of the Byzantine imperium. In the fourteenth century, the knowledge was transported back to Italy, where it soon became the most used technique of painting on walls [4], known through the works of Giotto and his followers and precisely described by Cennino Cennini, which talks about it as: »il più dolce e il piú vago lavorare che sia« (Cennini: III, 67) [5–7]. This is why fresco buono is generally linked to Italian Trecento and Quattrocento art; however, well before 1400 it spread around Europe through visiting artists from abroad or by Italian itinerant painters who started to look for work outside their homeland. On the other hand, a secco and lime techniques (the latter one believed to have developed toward the end of the Roman Empire, in a popular and provincial art as a cheap and fast substitute for the real a fresco [8, 9]) are generally more characteristic for Central and Northern Europe, probably due to the weather conditions – colder and more humid as in the South. Already the monk Theophilus from the twelfth century writes in his treatise about painting on a fresh layer of lime (I, 15, 16) [7, 10–12]. If there was really a barrier between northern and southern painting techniques, it surely disappeared with itinerant artists, who did not only spread the style along their way but also the manners of painting and the selection of materials. The knowledge from both directions merged and, as consequence, artists started to combine two or even all three painting techniques in one single work. The choice of a technique can reveal the dexterity of the painter – better he was, larger portion of a scene or a figure he managed to carry out a fresco, while the mortar was still fresh. But if the artist was slow realizing preliminary works, the plaster started to dry before the major part of the scene was carried out, so bigger portion had to be finished a secco or on lime wash. In general, the principal painting technique in Central Europe was fresco buono, but some modeling and last details were applied a secco. Rarely lime technique as the principal technique can be found, probably due to its instability (it peals easily from the wall). Also, the painter normally carried out the protagonists first and the secondary figures afterward, so many times he had to paint the latter ones on already dry mortar. He could also decide to paint as the first step the faces, hands, and bodies and only then the draperies, architecture, and backgrounds; therefore they could be made a secco in comparison to a fresco made flesh tones. Parts painted on a still wet surface can still be very well conserved today, while those carried out on a dry support can be completely lost.

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

617

The basic principle of fresco buono is application of the colors on a fresh plaster, made of binder (lime) and aggregate (sand, crushed marble or limestone, sometimes also other additives as crashed bricks, straw, etc.). The lime in the plaster serves as the binder for the aggregate and for the colors applied on plaster surface. For this purpose, slaked lime is used (Ca(OH)2), which is obtained from quick lime (CaO) by adding water, while quick lime is produced from lime rock/calcium carbonate (CaCO3) by heating it over 800
C. Slaked lime tends to convert back to calcium carbonate by drying – once exposed to the air, the water (H2O) evaporates and CO2 from the air binds again to the material which converts back to CaCO3, in the process of carbonatization (Fig. 22.2). During this process, the liquid slaked lime in the plaster tends to transport upward to the open-air surface and by this, surrounds the pigment grains. Once dried and converted in calcium carbonate, also pigment grains are captured between CaCO3 particles, and the color layer turns into a very solid surface [4, 8, 9, 13, 14]. Actually, it is not “glued” to the plaster surface, but it becomes one with the dried plaster. Therefore, fresco buono painting is the most resistant painting technique, because – if executed correctly – the colors do not peel off the surface. At the same time, the surface is characterized by sparkling crystals, formed through the carbonatization. In order to be able to paint on a humid surface, the plaster should be put on wall in portions, planned to be painted before the process of carbonization finishes. These portions are called giornata, a daily work, which in fact is limited only to several (3–5) h, depending on more or less humid climate. The artist should plan his work precisely and divide his painting in daily portions, in order to complete most of it on a still humid support, while the containing lime still has the binding power for pigments. To make the drying process longer, artists could use two or more plaster layers, each one smoother and with more lime. Many times, two layers were applied, sometimes even more. The penultimate layer is known as arriccio, while the last one, applied in daily portions, intonaco. Arriccio is rougher, it contains less lime and more sand with bigger granulation, while intonaco has more lime to offer stronger binding power, smaller sand grains, and sometimes crushed marble or lime-rock (both also CaCO3) instead of plain sand. This way, the plaster is white and very suitable for a fresco painting [4, 5, 8, 9, 14]. Lime technique is a simplified version of fresco buono, where the lime from the lime wash serves as the support and the binder and follows the same process of carbonatization; there is just no aggregate added to the lime [4, 8, 9, 13, 14]. The inconvenience in comparison to the plaster is that lime wash, being much thinner, even if applied in several layers, dries out much faster as the plaster, and therefore the

Fig. 22.2 Schematically presented process of carbonatization

618

A. Kriznar

time for painting on a fresh surface is much shorter; it can, nevertheless, easily be refreshed with another lime wash. As it is thinner as plaster, its binding strength is not high; therefore pigments should be also mixed with lime water or lime milk before their application on the support. The lime wash can easily peel off the wall, taking with it also color layers, that is, the painting itself. The quality of the painting depends on the quality of the lime, which is determined by the raw calcite material, its possible impurities, and its preparation. On the other hand, painting on a dry mortar has the advantage of unlimited painting time and a much wider pigment palette, as it will be explained in the next paragraph. However, pigments have to be mixed with a different, stronger binder than lime, normally an organic one, in order to be applied on the support. Among the most used ones are casein, animal glue, or egg yolk [4, 7, 8, 13, 14]. The painting on a dry plaster offers the possibility of several overlaid color layers, so the transitions between colors, lights, and shades can be easier to achieve. Nevertheless, and as the most important difference to fresco buono in terms of conservation, such color layers are just “glued” to the painting surface and do not form part of it; therefore, they can be peeled off or pulverized due to the disintegration of organic materials, etc., which leads to the loss of a secco color layers [4, 8, 13, 15].

22.1.2 Pigments and Binders Only limited number of pigments are suitable for painting on a fresh plaster, those that are stable in humid and alkaline environment: natural inorganic earths and minerals and some organic blacks [4, 5, 8, 9, 15, 16]. Synthetic pigments and most of the organic ones are more sensitive and can degrade, changing or even losing their color, as already warned by Cennini [5, 6] and can be studied in other literature [4, 8, 9, 16–18]. The palette for a fresco and lime painting consists, therefore, of lime white (Ca(OH)2) or bianco san Giovanni (CaCO3), when properly prepared [4, 5, 8], yellow (Fe(OH)3) and red (Fe2O3) ochres and clays, brown umber (Fe2O3(H2O) + MnO2(nH2O) + Al2O3), green earth (K[(Al,FeIII),(FeII,Mg] (AlSi3,Si4)O10(OH)2), minerals as azurite (2CuCO3Cu(OH)2), ultramarine (Na810Al6Si6O24S2–4), malachite (CuCO3.Cu(OH)2), cinnabar, and vermilion, its artificial version, with the same chemical composition (HgS). For black colors, mostly organic pigments were used due to their intense hue and much better stability as other organic colorants; mediaeval artists used mostly carbon, lamp, or vine blacks (all characterized by C). Even if these pigments are generally quite stable, they can suffer changes due to humidity, heat (e.g., fire), or air pollution. For example, the degradation of blue azurite seems to be related to pH and grain size. Due to humidity and wall moisture (with chloride salts), they can turn to green malachite or atacamite/ paratacamite (Cu2Cl(OH)3). Chloride ions from various sources can form black copper oxides (CuO), which can be also caused by exposure of the pigment to high heat (fire) in presence of alkali. The same blackening (formation of CuO) can occur with chemically similar malachite, which can, in fresco painting, also discolor due to the alkaline pH of the lime, especially when the particle size is small

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

619

[3, 8, 17]. Cinnabar can also turn black, changing to black metacinnabar (ß-HgS) [4, 16, 17]; therefore it was not considered a stable pigment for a fresco painting due to it sensitivity to humidity and alkali; still it was often used, especially in its cheaper artificial version. On the other hand, synthetic pigments used in mediaeval mural paintings are few, since not many have been yet known, and even those were unstable in humid and alkaline environment. Nevertheless, artists still used them due to their high covering power and intense brilliant colors, as it is a characteristic for all lead-based pigments: lead white (2PbCO3Pb(OH)2), yellow massicot and litharge (PbO) or their improved and more stable version of lead-tin yellow, type I (Pb2SnO4) or type 2 (Pb(Sn,Si)O3), and orange-red minium (Pb3O4). They had to be applied a secco with organic binders; still in most cases they suffered degradation (if not mixed with other pigments). They generally darken – turn brown or black – which is attributed to the formation of PbO2 due to the exposure to oxidizing agents or to the microbiological activity or to the formation of PbS under the reaction with H2S in atmospheric and biological agents. The presence of sulfuric pollutants can cause formation of lead sulfide (PbS) or lead sulfate (PbSO4) [3, 17]. Due to all these reactions, in many mural paintings, black areas can be observed, which have clearly been originally of different color, blue, green, white, yellow, or red (Fig. 22.3). On the other hand, painting on a dry surface does not suffer of alkali and moisture danger from the fresh plaster. All pigments can be basically used, including all inorganic, all organic, natural, and synthetic ones. They are too many to mention all, but there is a lot of literature written on the subject [8, 9, 13–18]. Nevertheless, the interior of a mediaeval building (church, palace) still tends to be humid, so organic materials are still not stable and can degrade, as well as organic binders used for painting on a dry surface. Egg yolk, casein, animal glue, or even oil can be chosen, but they all suffer different degradation processes [4, 8, 9, 13, 15]. Among them, oil is the worst selection, because the dried oil film does not let the wall to “breath,” which leads to the detachment and destruction of paint layers. Organic materials are easily attacked by microorganism; when binders degrade and slowly disappear, pigments have nothing else to bind them together nor to the wall, so the color layers start to detach, pulverize, and fall off.

22.1.3 Painting Procedures The working procedure applied by an artist can be his signature. Therefore, to study it is a very important part in understanding (a) an artwork, (b) the development of an artist, and (c) his influence on his disciples/followers. His selection of materials and painting techniques is a deciding factor in conservation of his work. When painting a fresco and applying at least two plasters, working by the system of small giornatas and using only suitable pigments, his paintings could last forever. To carry out a generally complex mural painting – especially if of large dimensions, when covering the entire interior – precise organization is necessary, starting with previous preparatory drawings (on paper or other available support), material selection and

620

A. Kriznar

Fig. 22.3 Pigment degradation: (a) changing of blue azurite to green (para)tacamite, Velemèr, Hungary (1377/1378); (b) blackening of a Cu-based pigment (azurite or malachite), Šenkovec, Croatia (a.1390); (c) lead pigment blackening, Kobenz, Austria (1420–1430)

collection, as well as dividing the space in scenes. If two layers of plasters are to be applied, arriccio can serve as support for the first preparatory drawing, sinopia (generally in red) [4, 5, 8, 9, 13], which helps at organizing the space on the walls (Fig. 22.4a). Later on, it is permanently covered by fresh portions of intonaco. Only if the latter one is strongly damaged, sinopia can be seen again. Sinopia was a characteristic step for Italian painters of Trecento and Quattrocento; however in Central and Northern Europe we rarely find it, especially due to the general use of one fresh plaster layer only. Because of this fact, its use can point toward the Italian origin of an artist. Once a daily portion of intonaco is applied, the artist draws the corresponding predrawing, as called the preparatory drawing on the upper plaster layer. The predrawing could be carried out in yellow, red, black, green, or even pink color, which can also be a signature for an artist. When painting on a dry plaster, the painter could draw the entire composition at once. Preparatory drawings are very important part of the painting procedure; they are generally made by the main artist himself, while the final work is a contribution of a larger workshop. Therefore, these preparatory drawings are especially valuable, because they reveal the dexterity of the artist – did he have strong decisive brushstroke, did he made many mistakes and had to change the composition, was he drawing fast with only principal lines or did

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

621

Fig. 22.4 Examples of preparatory drawings. (a) Sinopia on ariccio, Sazdice, Slovakia (1370– 1380); (b) predrawing on intonaco, Selo, Slovenia (1400–1410)

he carry out every detail, etc. At studying wall paintings, these drawings can be difficult to find, because they were covered by color layers and final contours; however, sometimes the color fell off or the painter did not cover them precisely, so they can be easily observed (Fig. 22.4b). For this, we can use several noninvasive techniques that will be explained in the next paragraphs. Among preparatory works are also incisions and pouncing. In incisions for straight lines, for haloes, and for attributes as chalice, crown, knife, belt, key, etc., most of them were decorated with rounded or linear pounced forms, as rays of halos or crown diamonds. For straight lines (decorative borders, architecture), a ruler and a sharp objects were mostly used; however sometimes also a rope could be splashed on the fresh plaster to impress straight lines; the rope could be dipped in red color, so the line was better observed, and by coloring it, this procedure could be also used as well on a dry plaster. The next step was the application of local tones and underpaintings. Local tones are the basic colors for carnation, drapery, architecture, etc., which serve as the base for later modeling of lights and shades. On the other hand, underpaintings are uniform color layers that serve as the base for another uniform color layer painted with a pigment that is preferably applied a secco or its color needs to be intensified. Sometimes the expensive red cinnabar was underlaid with much cheaper minium or red ochre; azurite and malachite, grained very finally, were applied over a gray (veneda, characteristic for Northern European art) or reddish (morello, characteristic for Italian Trecento and Quattrocento art) underlayer to intensify their colors and to spend less of these expensive pigments [4, 7, 9, 13, 15]. Less skilled and slower painters found themselves at this point in front of already dried plaster and had to finish the entire modeling and final details a secco, which made the painting more sensitive and not so stable, leading to worse conservation state. On the other hand, good quality artists continued with the modeling of the carnations, drapery folds, shades, and lights still in fresco buono; therefore the color layers are much better preserved, and the reading and interpretation of the murals is much clearer even today. In the way of modeling, also the skill of the artist can be evaluated – did he combine thin and thick brushes for soft transitions between

622

A. Kriznar

lights and shades, or was his modeling rough and the figures schematic? Are his final contours thick and dark or do they merge with the rest of the figures and objects? Can we find different painting procedures in one mural cycle? All these features can be discovered by the closer look at the paintings, using several noninvasive techniques. With the results, it is possible to confirm the authorship of an artwork, see the artistic development of an artist when compared to other works attributed to him, or even identify several artistic hands in one workshop, besides understanding better the conservation state of a mural painting.

22.2

Objectives

As pointed out several times, degradation of wall paintings is a serious problem; therefore it should be stopped or at least slowed down, eliminating the causes as far as possible. For this purpose, the state of a mural should be well studied and understood. When damages are important, restoration interventions (retouches, reconstructions) have to be carried out, which should be based on previous analysis of materials applied. The information on materials and painting techniques should be also valuable for art historians, yet they offer a reliable data on an artist/workshop that should be taken into consideration while studying the style of artworks, their influences, and connections. In order to get to know the materials, painting techniques and procedures of wall paintings, their state of conservation, and possible degrading factors, paintings should be thoroughly studied. The detailed objectives of such approach are to obtain data on: (a) Characterization of supports (plasters), including their composition (binders and aggregates) (b) Number of plaster layers and application of giornatas (c) Possible use of lime wash (d) Identification of pigments (inorganic, organic) and their possible degradation (e) Identification of binders (inorganic, organic) (f) Sequence of color layers (g) Color modeling and brushes in use (thin, thick) (h) Painting process from the preparatory work (under-drawings, incisions, pouncing, underpaintings, local colors) to the final color modeling (shades, highlights, final contour) (i) Chemical, physical, and biological agents that can degrade the painting and state of the conservation of the artwork

22.3

Methods

To obtain all these data, invasive and noninvasive analyses can be carried out. Invasive (or destructive) are considered in those which need a sample of plaster, pigment, or color layer to be analyzed with different analytical techniques (OM,

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

623

SEM-EDX, XRF, XRD, Raman, FTIR), normally in the laboratory with very precise analytical equipment. They can offer a precise information on material analyzed, especially if several of them can be combined. However, in many cases it is not possible or allowed to extract samples due to different reasons; therefore the information can be gathered by noninvasive (nondestructive) methods, which are becoming the first choice in material analysis, due to their quick development in the last years. They can still not offer as precise information as previously mentioned invasive techniques, but the data obtained is still very valuable. These analytical techniques are carried out in situ, without touching the surface of the artwork and therefore not damaging it in any way (techniques that apply strong radiation as X-rays are still under a question mark, not really knowing if some damages will be seen only after some decades [19]). The most useful methods for mural painting study are: (a) Precise examination by the naked eye (b) Use of different light sources (straight, raking light) (c) Digital microscopes and endoscopes for computer and/or mobile phones (d) UV and IR radiation (e) Portable X-ray fluorescence (XRF) spectroscopy (f) VIS reflectance spectrophotometry

22.3.1 Precise Examination by the Naked Eye with the Help of Raking Light A first look at a painting by the naked eye can offer first valuable information on the artwork. It determines the general conservation state, the color selection, and possible pigment changes. It is the most important step in this kind of research work, even if not really recognized as a method, and it should be carried out thoroughly. Enough light (daylight or artificial) is necessary for good observation, although not always possible – dark interiors, small windows, no electricity, etc. The first study reveals bigger or smaller damages in the plaster, detached areas, and, in consequence, missing parts of the painting; the surface can be covered with holes due to hammering (Fig. 22.1), carried out before the new plaster layer was applied over the paintings to hide them (generally due to the change of taste in art); crashing the crystalized surface made the binding between both plasters, the old and the new one, much stronger and long-lasting. We can also observe, if some color layers are missing, but the plaster is intact; they can be peeled off or pulverized due to degradation of organic binders, especially when painted a secco. The state of the conservation can point toward the selected painting technique by the artist; if the murals were carried out a fresco, color layers are generally still well preserved, even if there are plaster damages. On the other hand, colors applied with lime or a secco techniques tend to fall off. Under a good illumination, we can discover predrawings, incisions, underpainting, way of brushstrokes, and modeling; however, it is not always possible to

624

A. Kriznar

distinguish everything by the simple observation under the straight light, especially if we are searching for details. For a better observation, the use of raking light is very helpful, which can be pointed toward the surface from different angles. Big standing lamps can be applied, as used for restoration work, or only small flashlights can be chosen, which are even easier for changing the angle of the light striking the surface and, of course, to carry around. At a closer look, the plaster can be characterized: is it made of lime and sand as aggregate or crashed marble or lime wash were used (white surface), which points toward an Italian artist [4, 5, 9]. The yellowish color of the plaster reveals that the sand was not thoroughly washed and the lime not well prepared, so they still contain impurities, as iron oxides, clays, etc. By the naked eye, it can also be determined, whether the sand has big, small, or mixed granulation, if their form is round or angular, which reveals its provenance – from water or mine (Fig. 22.5). A closer look can show, if the plaster surface was well polished and prepared for the colors; by polishing fresh plaster, lime is pushed toward the surface, which makes the binding of the color layers with the plaster much stronger. If color layers are damaged and tend to peel off, it can be due to their application on lime wash, detected as white layer between the plaster and the colors. In some cases, also joints between giornatas can be discovered; skilled artists polished the joints very carefully and are practically invisible. They can be detected sometimes due to a slightly different color hue between two daily portions or with the help of the raking light (Fig. 22.6). We should look around the figures and scenes, along the decorative borders, trying to find a logical way of dividing a scene or a cycle. Less skilled artists did not make precise joints and are easy to discover. Further observation can reveal incisions and pouncing; the use of a ruler (Fig. 22.7a) or an impressed rope may be dipped in red color (Fig. 22.7b, 6c) for straight lines. The form and number of incisions and pouncing in a fresh plaster is also an important signature for an artist or his workshop and lack of them can as well

Fig. 22.5 Plaster observation: (a) big dark and angular sand grains spread in lime are easily observed under the daylight. The surface is not well polished and stays rough. Ofenbach, Austria (1415–1420); (b) uneven surface is better appreciated under the raking light, Kobenz, Austria (1420–1430)

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

625

Fig. 22.6 Giornata. Contact between two plaster portions can be sometimes well observed. Ľuboreč, Slovakia (a. 1385)

Fig. 22.7 Detection of (a) fine incisions in the raking light, Chyžne, Slovakia (fourteenth/fifteenth century); (b) impressed rope in fresh mortar, dipped in red color, under raking light, Martjanci, Slovenia (after 1492); (c) impressed rope dipped in red color on dry plaster, Chyžne

point toward the painting on already dry plaster. To find predrawings can be a difficult task using only the naked eye; in many cases they are hidden under the color layers and a special attention should be put to details as feet, hands, and drapery folds in order to discover, if a predrawing line is somewhere still visible. Having in mind that practically any color could have been used, one must be careful to distinguish a predrawing from a final contour. Nevertheless, some paintings reveal predrawings easily, as can be observed on the Fig. 22.4b.

626

A. Kriznar

By daylight and even better under the raking light, also underpaintings can be distinguished, especially gray or red under azurite or malachite; in many cases, these mineral pigments fell off, having been applied a secco, and so only the gray or reddish underpainting still remained. The modeling of colors, transitions between lights and shades, the form of brushstrokes, the final execution of the painting, and its style can as well be determined by the closer look. With the use of raking light, playing with the inclination of the light source, the sequence of color layers can sometimes be distinguished, observing better if lights were applied before or after the shades, how the carnation was constructed, were the color layers thin or thick, etc. (Fig. 22.8). In a fresco painting, this is easier to observe due to fewer color layers applied on a fresh mortar, while on a dry plaster possible more layers can distort the understanding of the color sequence. In this case, the area should be studied under magnification (loupe, digital microscope). As mentioned before, the first idea of pigment selection can be made during the observation, especially if we can deduce from previous steps that the painting is probably a fresco or lime one. This statement already narrows down the pigment palette; however, even if the painting is a secco, some pigments can be distinguished

Fig. 22.8 Modeling and color layer sequence of a foot observed under the (a) normal and (b) raking light; (c) thick color layers, following upward: white, yellow, orange, Turnišče, Slovenia (1383); (d) color layer sequence: yellow, red, black, and wide brush strokes, Rust, Austria (1400–1410)

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

627

according to their color. For example, green earth can be easily differentiated from malachite, having a darker tone and being generally applied in more than one layer due to its transparency. Therefore, it tends to peel off. Azurite has a characteristic bluish-greenish color (the intensity varies depending on how finely ground it is and if it is underpainted), but it can start degrading to green, which points toward the use of this pigment. Even clearer are the strange dark/black parts that point toward the use of lead and copper pigments that degraded (Fig. 22.3). By the eye inspection, of course, also most of later interventions and retouches can be located, especially if they were carried out using tratteggio technique. Many intervention treatments in the history were not documented, so this information also contributes to the knowledge on the state of the paintings today and their future treatment. This precise observation is possible, if the paintings are reachable from the floor, ladder, or scaffold (generally only possible when restoration works are already going on). But in bigger churches, the paintings can be too high on the wall or on the ceiling to be able to carry out such assessment. In these cases, good binoculars are necessary or a photographic camera which documents the state of the paintings in high resolution; photographs can be later studied on a computer. Another problem is darkness in some remote churches, where no electricity was yet wired. The use of light sources (lamps) connected to an external portable battery can solve such problem.

22.3.2 Digital Microscopes and Endoscopes Our eye is not able to see details or enter hidden places as cracks or plaster detachments. Sometimes a magnification is needed to clarify the doubts about the plaster composition, existence of lime wash, the sequence of color layers, etc. For this purpose, a simple magnifying glass can be used; however today several digital tools exist, small to carry around and not difficult to handle. They allow us also to record the images and study them later on.

USB Digital Microscope Small portable microscopes have been used for some years now to study the surface of the objects, in the laboratory or in situ. Their best advantage is that they are quite small (ca. 10 cm) and easy to handle. Today, there are already many on the market, but for artworks research not all are suitable. A chosen digital microscope should offer high magnification and enough resolution to get good quality pictures that can also be digitally magnified. It is also very convenient, if it includes UV and IR lights. It can be connected to a laptop through USB port, and the images are observed directly on the laptop screen. Each image can be recorded with a click on the microscope (it has a small very touch-sensitive button on the upper side) or on the record button in the computer program. One of the best at the moment is Dino-Lite Pro 2 digital microscope, model AM413T-I2V which has the possibility to switch between the white LED light, UV (~395 nm), and IRR (~940 nm) light – generally from the program on the computer, Dino Capture 2.0. Its objective has 20–200x magnification and the resolution is 1.3 MP (1280  1024), which offers enough

628

A. Kriznar

quality images to be studied later on. There exist also other similar microscopes that have much higher magnification and up to 5MP resolution, but they do not include IR and UV lights. Under the normal white light, the surface can be well observed. If studying the plaster, the magnification helps us see the aggregate particles (size, color) or possible lime wash laid over it. It can reveal the sequence of color layers, many times difficult to observe by the naked eye only. This is very useful especially when no samples can be extracted to be studied as cross sections under an optical microscope. The magnification is also important to observe small damages and holes, “craters,” in the plaster and color layer, caused by the impact of salts in the wall. With UV and IR light, the small observed area can be studied also in search for later interventions, and – in some cases – it can help us identify some pigments (Fig. 22.9). As studying wall paintings, there can be several problems by using this kind of digital microscope. First, if the paintings are too high, they cannot be reached. Rarely there is a possibility of scaffolding or even a ladder. Besides, if one works alone, it is very difficult or even impossible to handle the laptop and the microscope at the same time. A support for the laptop that can be moved along the wall can help, in order to manage the microscope with free hands. Luckily, Dino developed a new wireless version, which can be switched to USB easily by changing the adaptor. The best model for now is WT4515ZT Digital Wireless AMT polarizing microscope, with 1.3 MP resolution and 20–220x optical magnification, but no UVand IR lights, and the images captured can only be observed on the computer screen through the software provided. A better solution for studying mural paintings are digital microscopes (still few on the market), which have already attached a small display; therefore no laptop is necessary and in situ work is much easier. Every image can be observed on the display and saved to the SD card. A good example is Toolcraft Portable Digital Microscope that offers 30–220x optical magnification with additional digital zoom if required, has 5MP resolution, and has an LCD built-in screen and a polarizer. The Li-ion battery makes possible to work up to 4 h. It has very good characteristics for surface analysis, especially for restorers; however it lacks UV and IRR light for a deeper insight.

Fig. 22.9 Wall painting surface studied by Dino-Lite microscope, under (a) normal, (b) UV, and (c) IR light. Different color layers and their texture can be distinguished, while dark spots can reveal later small retouches. Gombašek, Slovakia (fourteenth century)

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

629

Digital Microscope for Smartphones To make the magnification of the surface even easier, there are also very small microscopes (about 1–2 cm) that can be placed directly over the camera lens of a smartphone. As it is already usual to have our phones with us most of the time, a digital microscope is just a small gadget to carry with and very easy to put on. The phone with the digital microscope can be placed directly next to the surface (without touching it, so that color layers don’t get damaged) and the image is displayed on the phone screen, using the app that corresponds to the microscope. Generally, it just does not offer adjustable magnification, as the USB digital microscopes do. There are several on the market, for iPhones and Androids of different kind, and the price range is quite wide as is their resolution. Two most general ones which can be put on practically every phone are Earldom and Nurugo. As the first one is more stable once coupled to the phone, it offers only 200x magnification, while the second one has the possibility of 400x magnification and can be, therefore, more helpful distinguishing color layers. It is, nevertheless, difficult to keep it in place, being attached to the phone only with a plastic handle which is not very stable. All smartphone microscopes use the backlight of the camera, so the illumination of the studied area is very good. A picture can be taken from the phone’s screen (through the corresponding app) by touching it and saving it on a phone memory card. It is the easiest way to study the murals surface closer, especially if we work alone. With this simple tool, plaster surface and sand grains can be well observed, as well as possible lime wash, the sequence of color layers, or their detachment (Fig. 22.10). Digital Endoscope Wall paintings can be damaged by deep cracks, holes, or plaster detachments. If the damages are deep, it is not possible to see by the naked eye, what is inside or how far away does a crack/detachment go, which is an important information for evaluating the conservation state of the artwork and plan its treatment. In this case, a very handy tool exists, a digital endoscope, which can be easily plugged in a smartphone (Fig. 22.11) or with the USB adaptor to a computer. Using the app or software, the image is displayed on the phone screen or on the computer screen. Generally, endoscopes come with a long cable, so it is easy to go deep into a crack. Around the endoscope camera, small LED lights are placed to illuminate the dark areas, which enable us to get a clear image. Pictures can be taken or even a video recorded, which is saved on the smartphone memory card and can be easily downloaded on the computer. As with digital microscopes, also with endoscopes, it is important to know the resolution. There are several on the market (e.g., CameraFi, Depstech, Fantronics), most of them have 2.0 MP resolution and can be used for Android, iPhone, Mac, or Windows. Three very useful accessories can come with the endoscope: a small side sight mirror, a magnet, and a hook, all of which can be coupled on the endoscope and used in a crack or hole. (b) endoscope coupled to a smartphone, studying a crack and a detachment in a wall

630

A. Kriznar

Fig. 22.10 Painting surface observed under the digital microscope for smartphones. (a) Microscope attached to the phone; (b) sand grain; (c) red and green color layers on plaster. Šenkovec, Croatia (ar. 1390)

Fig. 22.11 (a) Detachment between two plasters can be observed on the endoscope digital image

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

631

22.3.3 Ultraviolet and Infrared Radiation For a better study of the surface and pigments applied, besides natural light, also UV or IR radiation are often used in artworks characterization, both of them offering an important information.

UV Radiation Ultraviolet light radiation is based on a different wavelength as the visible light, just above it; therefore, UV energy is lower than that of the visible light. When UV rays interact with a material (as all electromagnetic waves), one of the reactions produced is UV fluorescence. Every element irradiates a fluorescence that is a characteristic for it, which can identify the irradiated material [20– 22]. The UV light can be obtained with the so-called Wood’s lamps, which can have different size and power. The bigger ones can be placed in front of the surface, while the smaller, handheld ones can be moved easily along the surface to observe the fluorescence. The radiation caused by UV fluorescence can be seen by the naked eye and is displayed with different colors from light green, yellow, light orange, to brown, depending on the material that irradiates it [23]. Such image can also be photographed and documented by a normal camera. This method is very important especially for discovering later interventions, having in mind that aged materials fluorescence differently as new ones (Fig. 22.12). In

Fig. 22.12 Mural painting observed under UV light. On the image, many later interventions can be observed in dark brown and bright yellowish areas. The allegory of the trade of crafts and sciences by Andrej Janez Herrlein (1786), Grubar Palace – Archive RS, Ljubljana, Slovenia. (Foto: Tine Benedik, Archive ZVKDS, RC, Slovenia)

632

A. Kriznar

this way, later retouches are observed as grayish, brownish areas in contrast to lighter fluorescence of older pigments. With this method, also the presence of organic materials or zinc pigments (in use from the end of the nineteenth century) can be detected, which under UV light get a strong yellow glow. Although this technique is very useful, applying it on wall paintings is not so easy and, actually, does not always offer good results due to already strong fluorescence of the lime, gypsum, and pigments bind with lime. On the other side, for good UV results, we should be working in the dark; however in churches, this is generally not possible due to windows without shutter; we might have a possibility to work at night, though. In spite of these difficulties, this method is good to discover almost lost pigments, parts carried out a secco (due to a fluorescent organic binder) and later interventions [8].

IR Light Wavelength of infrared light is just under the visible light; therefore, its energy is higher, and the radiation can penetrate deeper in the material. It can discover what is hidden under the upper color layers [21, 22]. Results can be recorded with IR photography which uses special filters that eliminate the spectrum of the visible light. The photography can be black-white or in color, but the colors are different as the real ones. This radiation can discover preparatory drawings (especially if they are carried out in black, while red color cannot be detected by IR), can identify some pigments with a characteristic grayish hue, and therefore can also reveal some later interventions; these might have been carried out with the same color, but different chemical composition, which is easily revealed under the IR radiation. Infrared waves are used also for IR reflectography [20, 21, 24, 25], which generally applies radiation with higher energy, so the waves can penetrate even deeper. This way, better results especially for preparatory drawings or even over-paintings and composition changings can be expected. A special IRR camera records the radiation reflected from the surface, and since it is connected to a computer, the images are recorded directly to the hard drive and can be then assembled to a bigger image. Several programs are available for that purpose. The application of IR radiation technique is not very common in the study of mural paintings; better results can be expected analyzing panel and canvas paintings. The problem can be the electricity in a church, needed for the camera, for the computer, and principally for halogen lights, source of the IR radiation. Also, the position of the murals, often very high, makes very difficult, if not impossible to reach the paintings. Nevertheless, through IRR some pigments or later retouches can be identified. However, discovering a preparatory drawing, which is the principal use of this method for artworks on other supports, is very optimistic – as described above, predrawings are mostly carried out in red, which is “invisible” for IRR, but also yellow is not well readable. It could give good results for black predrawings, but in murals they are not so common. In any case, sinopia cannot be seen, because it is hidden under a layer of intonaco, which is too thick to be crossed by IR radiation.

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

633

22.3.4 Portable X-Ray Fluorescence (XRF) Spectroscopy According to their wavelength, X-rays are situated between UV and gamma rays and are characterized by short waves and high energy. Therefore, they are able to penetrate deeper into the material as previously mentioned visible, UV, and IR waves, which makes them very useful in the study of many objects, also in cultural heritage. The basis of all X-ray analytical techniques is the interaction between the rays and the irradiated material, which originates different reactions [20, 21, 26, 27]. Only a small part of this electromagnetic waves go through the material without changing their energy or direction; most of them are absorbed, emitted, or reflected or they originate other processes. The name of the method is the reaction which we measure and among the most useful ones in cultural heritage are X-ray fluorescence spectroscopy, X-ray diffraction, and radiography. X-ray fluorescence technique has been used in analysis of artworks intensively over years now [22, 27, 28]. There are many instruments on the market, some commercial and others “home-made” by several research groups. They are composed of X-ray tubes with different anode wires (W, Rh, Ag, Mo, W, Cu, etc.) and detectors (PIN diode, SDD, etc.), whose energy and intensity can vary. They might need the application of a collimator to reduce the beam size or of a filter (Al, etc.) to supreme peaks from the anode (e.g., W peaks “cover” the energy area of Fe, Cu, and Zn, important in pigment identification; therefore they should be eliminated from the spectrum). Some instruments can be used only in the controlled environment of a laboratory, while others, generally smaller, can be transported for in situ measurements, to museums, galleries, restoration workshops, etc. The transportability/portability of an equipment is of high priority for the study of artworks – they do not need to be removed and the danger of possible damage is much smaller. For analysis of wall paintings, only portable instruments can be applied, because the artworks are bound with the architecture and cannot be moved, except when detached. Therefore, the measurements should be carried out in situ (Fig. 22.13a), and the equipment should be easily movable around the space, having in mind that a mural cycle can cover the entire nave and/or presbytery of the church. Therefore, the smallest it is, the better. The easiest to operate is a handheld XRF, which was originally applied mostly for metals and earth crust measurements, but during the last years, it developed to one of the most used/useful methods for archaeometry analysis. Several handheld instruments are already on the market, however, among the most used and known one in the field of cultural heritage is the series of Bruker Tracer μ-XRF Spectrometer ArtTAX. This series is equipped with a live-time spectral display which shows the results at once, and it can work with Li-ion battery, which makes the work in situ much easier and effective. It can also be connected to a laptop, and analyses are carried out from the corresponding software program. After all optical noninvasive techniques presented previously, this is the method, which offers semi-quantitative elemental analysis of the material studied [21, 27, 28]. In case of mural paintings, we can measure pigments and support (plaster). The technique is based on an optical phenomenon, fluorescence, which occurs when

634

A. Kriznar

Fig. 22.13 In situ analysis of mural paintings in the Holy Trinity Church in Velemèr, Hungary (1377/1378), by (a) portable μ-XRF and (b) portable VIS spectrophotometry

an electron with higher energy releases part of its energy, replacing a void space in a lower energy orbit in an atom, caused by X-ray radiation. The energy and the length of these waves depend on the atomic number of chemical element(s) in the irradiated material and are characteristic for each element from the periodic table. By this, it is possible to identify chemical elements that compose the analyzed material, but not its chemical composition. Knowing chemical elements, material studied can be identified in situ, without sample extraction and generally already during the measurement itself, which can take only some seconds or up to several minutes, depending on the work premises. In case of wall paintings, pigments on the base of their characteristic elements can be identified: lime white (Ca), ochers and clays (Fe), umber (Mn, Fe), green earth (Mg, Al, Si, K, Fe), malachite and azurite (Cu), lead pigments (Pb), and cinnabar (Hg), to mention the most relevant ones. Nevertheless, there are some limitations. XRF can identify only chemical elements with Z > 14 (heavier than Si), what leaves all organic materials out of its reach, as well as the expensive mineral ultramarine, composed by only light elements (Na, Al, Si, O, S). On the other hand, pigments that are characterized by the same chemical element can also not be distinguished, except by their color – if they did not suffer some chemical changes. For example, with XRF we cannot determine different copper-based green pigments, and there are many used through the history. Nevertheless, when analyzing mediaeval wall paintings, the only used copper green pigment is used to be malachite, which helps at its identification (or at least hypothesis). Also, all Pb-based pigments can be identified only as lead pigments, but with XRF it cannot be concluded, whether it is lead white, massicote, litharge, or minium, except if we clearly distinguish the color that was measured. The identification gets even more complicated, if the pigments have darkened, as it is often the case precisely with Pb-based ones. Only lead-tin yellow can be properly identified, when finding Pb and Sn in the same analyzed area. For the lime wash and the plaster, XRF principally detects lime (Ca), sometimes also Mg, Al, Si, K, and Fe, often only as traces that reveal plaster impurities (dolomite, quartz, iron oxides). Trace elements can be important also to

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

635

distinguish between two chemically very similar pigments (with the same principal chemical element) or even help us to find the provenance of a pigment, which is actually very rare. In case of the Holy Trinity Church in Velemèr (Hungary), no samples could have been taken for its study due to the high historical and artistic value of the paintings, carried out by Johannes Aquila (1377/1378) [29, 30]. Therefore, only XRF was applied, using the ArtTAX Tracer III-SD instrument, with Rh tube (max 30 μA at 40 kV) and 10 mm2 X-Flash Detector (145 eV). In the nave and the presbytery of the church, painted entirely, about 130 points of different colors were taken, using always the same working conditions: voltage ¼ 40 kV, current ¼ 15 μA, and time ¼ 20s. The short measurement time should be pointed out, as it makes possible to work quickly and therefore, analyze many selected points. The only limitations were the time (only some hours permission) and high, inaccessible parts of the painting, which could not be reached. Plaster and all different colors with their lighter/darker tonalities were studied, as well as dark/black areas that pointed toward a pigment degradation. Comparing the spectra, not only inorganic pigments can be identified, but also their mixtures, for example, if the lights were obtained by adding a white pigment or just diluting the basic color, or if shades were created by adding dark (brown, black) pigment. In this case study, all inorganic pigments were identified, lead pigment changes (high Pb presence in several black areas), and degradation of azurite to (para) tacamite (presence of Cl in the changed green area, but not in the still original blue one) were confirmed. Taking into consideration also trace elements, it was possible to distinguish between the red and violet pigments, both are characterized by strong Fe peaks, but the red ochre contains also traces of K, Ti, and Zn, hardly present in the violet one (Fig. 22.14b). On the other hand, traces of As and Zn in green malachite (high Cu peaks) showed possible provenance of the mineral in Slovak Ľubietova copper mine, as discovered in previous studies [31] (Fig. 22.14a).

Fig. 22.14 (a) The presence of As and Zn impurities in the Cu-based green pigment points toward the origin of the pigment in Slovakia; (b) analysis of the red and violet color shows important difference in trace elements

636

A. Kriznar

The interpretation of the spectra can be complicated, having in mind that the beam transverses several layers, and in some cases, only hypothesis can be established. Nevertheless, in wall paintings, the understanding of the results is easier, as they usually do not have as many layers as panel or canvas painting. However, pigment mixtures could have been applied, and there is always high Ca intensity present. Calcium signals can come from the plaster, from the possible lime wash, from the lime binder used for pigments, or from lime white, added to other pigments, just to give the example of the complexity of the results obtained in wall paintings. Still open questions can be clarified, if some samples can be extracted after all, and studied as cross sections with other methods (OM, SEM-EDX, FTIR), which already enter the area of invasive techniques, not subject of this chapter. In this case, the previous application of XRF is also highly useful, because it narrows down the number of necessary samples and points toward the areas of interest. Samples are, therefore, extracted only from a few controlled areas, which are important for the conservation of the murals. In recent years, a portable combination of XRF con XRD was achieved [26, 32, 33], which can offer much better results, not only elemental from XRF but also compositional (crystalline structures) from XRD. Therefore, it is useful for inorganic and also organic materials, as well for the identification of degraded pigments, and those not possible to distinguish by XRF due to the same principal chemical element. Some instruments are still difficult to handle in situ due to their size and/or form, but some analysis of wall paintings have been already successfully carried out [34].

22.3.5 Portable VIS Spectrophotometry The perception of our eyes is not objective due to different reasons; therefore we cannot always distinguish a precise color (green-blue, orange-red, brown-black, etc.). When studying mural paintings, we must also consider possible dark interiors, high painting position, and, of course, pigment changes that have been mentioned previously. The perceived color depends on the pigment color, the binding medium, surface absorbency, texture of the surface, size of particles, underlayers, etc. For an accurate classification and matching of the colors in a painting, a useful technique is VIS spectrophotometry, which deals with the measurement of color [26, 35]. It can quantify the visible effects that are related to the coloring materials, in case of wall paintings, the pigments. VIS involves only surface measurement; therefore underlying paint layers cannot be analyzed. The color is a result of selective absorption/emission of light which is characteristic for each color, and this phenomenon can be measured with special equipment, VIS spectrophotometer. The technique is based on the transition of electrons between the outer orbits (energy levels) of an atom, whose energies are in the range of near IR, visible, and UV waves. A monochromator placed between a light source and the analyzed surface allows selection of the wavelength of the beam, while a detector records the photon intensity which can be translated to a spectroscopy curve. Technique based on transmission is not suitable

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

637

for mural paintings, because the light cannot go through the analyzed material. Therefore, for cultural heritage application, surface reflectance techniques have been developed, which use fiber optics (FORS – fiber optics reflectance spectrophotometry) and digital recording technologies, which can be used for different types of artworks as inks, colorants, and wall paintings [26, 35–38]. There is a large availability in the market, most of them are characterized by a very short measurement time needing just a few seconds for spectrum collection. Reflectance spectra help us to identify and classify colors, which is important for the complete understanding of the paintings (also from the iconographic point of view) and especially for an accurate intervention process and its restoration (reintegration with according colors). The color identification is based on the comparison of the obtained spectrum (curve) with reference material; therefore, a good database is necessary. Reflectance spectra can be used to distinguish between similar colors and hues by the identification of maximum peaks and inflection points, but these features can strongly vary, if we measure a mixture of two or even more pigments. This is why this technique gives good results for single pigments and colorants, but when measuring the mixtures (mostly the case in paintings and polichromy), the reflectance curve can shift and the material is hard to identify [36, 37]. For in situ analysis of the murals in Velemèr (Fig. 22.13b), the spectrophotometer SPM 100 (Gretag Imaging, AG) was applied. It measures the reflectance of visible light from 380 to 730 nm, with a spectral resolution of 10 nm. The surface of the sample was illuminated for 0.5 s using a small 2 W bulb with a spot size of 4 mm. The obtained spectra were compared with an internal database, elaborated in the research group Kunst- und Kulturgutanalyse (4.5) at the BAM (Bundesanstalt für Materialforschung und –prüfung, Berlin, Germany). The reflectance spectra were normalized and the first derivative was formed for better peak comparison. Several colors have been measured, in order to properly classify them, especially in the cases of doubt (green-blue; red-brown). Since most of them are mixtures of at least one pigment with lime binder, several obtained spectra were not clear, as is the case of measuring azurite (Fig. 22.15a) and its degradation into (para)tacamite. However, in the case of darkened colors, identified by XRF as lead based pigments, VIS results suggest the use of minium (Fig. 22.15b) and by this complemented the XRF results. As other in situ measurements, also the application of this method can be complicated due to the large size of an artwork, rough surface, and the inaccessible location of the color of interest. It is also complicated to keep the instrument very close to the vertical painting surface (1 mm) without touching it, especially when pressing it for each measurement.

22.4

Conclusions

Even if the so-called invasive techniques give generally better or more precise results, there is a large spectrum of noninvasive techniques that help us identify materials applied in mural paintings and their technique of execution. Several of them are very easy to use and the instruments are cheap and affordable, as is the case

638

A. Kriznar

Fig. 22.15 Analysis of murals in Velemèr by VIS spectrophotometry; spectra obtained are pointed by an arrow, compared to the reference material in color. (a) Measurement of the blue color (azurite) did not offer conclusive results, while (b) the spectrum of the red darkened area suggests the presence of minium. (reference database by Simon Steger, FB 4.5, BAM). Velemèr, Hungary

of digital microscopes and endoscopes. The application of UV and IRR techniques is useful; however in the case of in situ study for murals, they do not always offer good results due to the principal materials used in such techniques, as lime (strong fluorescence that can prevent to see other fluorescence) or the use of red pigments (not detectable for IRR). The use of XRF and VIS spectrophotometry offer one step forward in the analysis, making possible the identification of material on the basis of their chemical and optical characteristics. It must be, however, pointed out that the visible inspection by the naked eye is an essential first step for artwork studying, before any other equipment should be used, having in mind the variety of information that it can already be gathered on materials, painting technique, state of conservation, etc. Noninvasive analysis is important especially in cases where no samples can be taken, but it is also very welcome as the preliminary study before sample collection, reducing considerably their number. The knowledge on materials and techniques used by mediaeval artists should be of interest not only to conservators and restorers but also to art historians and everybody dealing with cultural heritage. Acknowledgments The author acknowledges (a) the Alexander von Humboldt Foundation (Bonn, Germany) for the scholarship for advanced researchers, which made possible to carry out an extense project on analysis of Central-Eastern European mediaeval wall paintings, (b) the BAM – Bundesanstalt für Materialforschung und -prüffung, group 4.5 Kunst- und Kulturgutanalyse for accepting her as guest researcher, making possible the realization of analyses of selected monuments, and (c) the Archive ZVKDS, RC of the Republic of Slovenia for the UV image.

References 1. Mora P (1974) Causes of deterioration of mural paintings. International Centre for the Study Of The Preservation and the Restoration of Cultural Property, Rome 2. Massari G (1971) Humidity in monuments. University of Rome, Faculty of Architecture; International centre for the study of the preservation and the restoration of cultural property, Rome

22

Noninvasive In Situ Analysis of Mediaeval Mural Paintings

639

3. Cocciato A, Moens L, Vandenabeele P (2017) On the stability of mediaeval inorganic pigments: a literature review of the effect of climate, material selection, biological activity, analysis and conservation treatments. Heritage Science 5(12):2–25. https://doi.org/10.1186/s40494-017-0125-6 4. Knoepfli A, Emmeneger O (1990) Wandmalerei bis zum Ende des Mittelalters. In: Knoepfli A, Emmeneger O, Koller M, Meyer A (eds) Reclams Handbuch der künstlerischen Techniken, vol 2. Philipp Reclam jun, Stuttgart, pp 7–212 5. Cennini C (1999) In: Serchi M (ed) Il Libro dell’arte, Firenze 6. Broecke L (2015) Cenino Cennini’s Il Libro dell’arte. Archetype Publications, London 7. Mazzè A (1998) La decorazione murale, Stucchi, affreschi, graffiti nella trattatistica. Palermo, Sao Paolo 8. Mora P, Mora L, Philippot P (2001) La conservazione delle Pitture Murali. Editrice Compositori, Bologna 9. Philippot P (1984) Die Wandmalerei, Entwicklung, Technik, Eigenart. Koehler & Amelang, Lepizig 10. Theophilus (1874) In: Ilg A (ed) Schedula diversarum artium. Wilhelm Braumüler, Wien 11. Theophilus (1979) In: Hawthorne JG, Smith CS (eds) On divers arts: the foremost medieval treatise on painting, glassmaking and metalworking. Dover, New York 12. Berger E (2009) Fresko- und Sgraffito-Technik nach ältern und neuern Quellen. Sändig Reprint, München 13. Eibner A (1970) Entwicklung und Werkstoffe der Wandmalerei vom Altertum bis zur Neuzeit. Sänding Reprint, München 14. Wehlte K (1962) Wandmalerei, Praktische Einfürung in Werkstoffe und Techniken. Otto Meier Verlag, Ravensburg 15. Brachert T (2001) Lexikon historischer Maltechniken, Quellen – Handwerk – Technologie – Alchemie. Callwey, München 16. Montagna G (1993) I pigmenti. Prontuario per l’arte e il restauro, Firenze 17. Fitzhugh EW, Feller RL, Roy A, Berrie B (eds) (2012) Artists’ pigments: a handbook of their history and characterisation. Archetype Publication, London 18. Eastaugh N, Walsh V, Chaplin T, Siddall R (2008) Pigment Compendium, a dictionary and optical microscopy of historical pigments. Elsevier, London, New York 19. Mantler M, Klikovits J (2004) Analysis of art objects and other delicate samples: is XRF really nondestructive? Powder Diffraction 19(1):16–19. https://doi.org/10.1154/1.1649962 20. Gómez ML (2000) La restauración, Examen científico aplicado a la conservación de obras de arte. Cátedra, Madrid. in Spanish 21. Matteini M, Moles A (1994) Scienza e restauro. Metodi d’indagine. Nardini, Firenze 22. Volpin S, Appolonia I (2002) Le análisis di laboratorio aplícate ai beni artistici policromi. Il Prato, Padova 23. Botticelli G (1992) Metodologia di restauro delle pitture murali. Centro Di, Firenze 24. Casini A, Lotti F, Picollo M, Stefani L, Buzzegoli E (1999) Image spectroscopy mapping technique for non-invasive analysis of paintings. Stud Conserv 44(1):29–48. https://doi.org/10. 2307/1506694 25. De Boer JRJVA (1968) Infrared reflectography: a method for the examination of paintings. Applied Optics 7(9):1711–1714. https://doi.org/10.1364/AO.7.001711 26. Artioli G (2010) Scientific methods and cultural heritage: an introduction to the application of materials science to archaeometry and conservation science. Oxford University Press, Oxford, New York 27. Albella JM, Cintas A, Miranda T, Serratosa JM (1993) Introducción a la ciencia de materiales. Consejo Superior de Investigaciones Científicas, Madrid. in Spanish 28. Seccaroni C, Moioli P (2004) Fluorescenza X: Prontuario per l’analisi XRF portatile applicata a superfici policrome. Nardini editore, Firenze 29. Radocsay D (1977) Wandgemälde im mittelalterlichen Ungarn. Corvina Verlag, Budapest 30. Höfler J, Balažic J (1992) Johannes Aquila. Pomurska založba, Murska Sobota 31. Hradil D, Hradilová J, Bezdička P, Švarcová S (2008) Provenance study of Gothic paintings from North-East Slovakia by handheld x-ray fluorescence, microscopy and x-ray microdiffraction. X-Ray Spectroscopy 37:376–382. https://doi.org/10.1002/xrs.1014

640

A. Kriznar

32. Gianoncelli A, Castaing J, Ortega L, Dooryhee E, Eveno M, Salomon J, Bordet P, Hodeau JL, Walter P (2007) A portable XRF-XRD instrument for in-situ analysis of cultural heritage objects. In: Townsend J (ed) Conservation science. Archetype Publications, London, pp 189– 194 33. Eveno M, Moignard B, Castaing J (2011) Portable apparatus for in situ X-ray diffraction and fluorescence analyses of artworks. Microsc Microanal 17(5):667–673. https://doi.org/10.1017/ S1431927611000201 34. Duran A, Castaing J, Walter P (2010) X-ray diffraction studies of Pompeian wall paintings using synchrotron radiation and dedicated laboratory-made systems. Applied Physics A 99: 333–340. https://doi.org/10.1007/s00339-010-5625-0 35. Bacci M (1995) Fibre optics applications to works of art. Sensors Actuators B Chem 29(1–3):190– 196. https://doi.org/10.1016/0925-4005(95)01682-1 36. Aceto M, Agostino A, Fenoglio G, Idone A, Gulmini M, Picollo M, Ricciardi P, Delaney JK (2014) Characterization of colourants on illuminated manuscripts by portable fibre optic UVvisible-NIR reflectance spectrophotometry. Anal Methods 6:1488–1500. https://doi.org/10. 1039/C3AY41904E 37. Oltrogge D (2008) The use of VIS spectroscopy in non destructive paint analysis potential and limits of the method for 19th and early-20th-century paintings. Museen Köln, Köln. OnlineEdition. http://www.museenkoeln.de/impressionismus 38. Bacci M, F. Baldini, R. Carlà, R. Linari, M. Picollo and B. Radicati: Color analysis of the Brancacci Chapel Frescoes, Appl. Spectrosc., 45 (1), pp. 26–31 (1991), https://doi.org/10.1366/ 0003702914337713 & Part II: Appl Spectrosc 47(4):399–402 (1993), https://doi.org/10.1366/ 0003702934335074

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

23

Anabelle Kriznar, Kilian Laclavetine, Francisco J. Ager, Claudia Caliri, Francesco Paolo Romano, and Miguel A´ngel Respaldiza

Contents 23.1 23.2 23.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives and Analytical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . UV-Induced Visible Fluorescence (UVF) Photography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1 Experimental Procedure and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

642 643 644 646

A. Kriznar (*) Departamento de Escultura e Historia de las Artes Plásticas, Facultad de Bellas Artes, Centro Nacional de Aceleradores, Universidad de Sevilla, Seville, Spain Department of Art History, Faculty of Arts, University of Ljubljana, Ljubljana, Slovenia e-mail: [email protected] K. Laclavetine Centro Nacional de Aceleradores, (Universidad de Sevilla-CSIC-J. Andalucía), Seville, Spain Departamento de Física Atómica, Molecular y Nuclear, Universidad de Sevilla, Seville, Spain Centre de Recherche et de Restauration des Musées de France (C2RMF), Paris, France Centre de Recherche sur la Conservation des Collections (CRCC), Paris, France F. J. Ager Centro Nacional de Aceleradores, (Universidad de Sevilla-CSIC-J. Andalucía), Seville, Spain Departamento de Física Aplicada I., Universidad de Sevilla, Seville, Spain e-mail: [email protected] C. Caliri · F. P. Romano INFN, Laboratori Nazionali del Sud, Catania, Italy CNR, Istituto Scienze del Patrimonio Culturale, Catania, Italy e-mail: [email protected]; [email protected] M. Á. Respaldiza Centro Nacional de Aceleradores, (Universidad de Sevilla-CSIC-J. Andalucía), Seville, Spain Departamento de Física Atómica, Molecular y Nuclear, Universidad de Sevilla, Seville, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_23

641

642

A. Kriznar et al.

23.4

648 648 654 658 660 664 664 668 669

Infrared Reflectography (IRR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4.1 Experimental Procedure and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4.2 Specific IR Reflectivity of Colorant Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.5 Portable X-Ray Fluorescence (XRF) Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.5.1 Experimental Procedure and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.6 Macro X-Ray Fluorescence (MA-XRF) Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.6.1 Experimental Procedure and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

Analysis of artworks in museums and galleries has become one of the most important features in conservation and restoration, and even in authentication and attribution, when doubtful. Materials applied by the artists, as well as painting techniques and procedures, have strong influence on the conservation state of an artwork, besides other historical and environmental agents. Many noninvasive techniques can be applied in order to know all these aspects, which should be applied as the first step in artworks analysis, prior to other invasive methods, which should only complement them, obtaining some more detailed results. Among the most used ones are UV-induced visible fluorescence (UVF) photography, infrared reflectography (IRR), and X-ray fluorescence (XRF) spectroscopy, with the latest developments such as macro X-ray fluorescence (MA-XRF) spectroscopy. UVF photography can show later interventions, IRR can reveal preparatory drawings and pentimenti, while XRF spectroscopy is used for elemental analysis of the materials applied. MA-XRF spectroscopy shows the distribution of chemical elements (and therefore pigments and possible changes in the composition) throughout the entire painting surface. All these techniques can be used directly in situ, in the exhibition hall or in the restoration workshop and without direct contact with the surface of the artwork. For security reasons, the study using X-ray must be done in an area closed to the public. Keywords

Non-invasive techniques · paintings · pigments · retouches · preparatory drawings · material analysis · UVF · IRR · XRF · MA-XRF

23.1

Introduction

Art collections in museums and galleries are nowadays usually well monitored regarding temperature, humidity, and light [1]; nevertheless, artworks can show signs of damage due to many reasons [2, 3]. This occurs not only because of the natural degradation of materials but also due to previous and generally not suitable storage and exhibition conditions during the lifetime of an artwork. Therefore, restoration interventions must be planned, in order to return the artwork to its best condition and to eliminate any possible environmental danger; besides the basic atmosphere agents, also air pollution can be very harmful, causing discoloration of some pigments, as well

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

643

as biological activity that attacks organic materials, such as colorants, binders, and varnishes. Another factor to have in mind in museums and galleries are also not properly informed and educated visitors who touch artworks, or, in an extreme case, vandalize them. A good conservation and restoration intervention can be carried out only when the professionals know the materials, techniques, and environmental circumstances of an artwork to be restored. Therefore, analysis of materials must be selected as the first step in saving a painting or a sculpture. Supports, pigments and colorants, binders, and varnishes can be studied with different techniques, many of them noninvasive. The selection varies on the accessibility to instruments by every institution; many museums already have some of the basic equipment, while others chose to collaborate with scientific research groups that usually handle the necessary instrumentation. In this chapter, the authors concentrate on paintings, being an important and complex part of many collections. In order to obtain good analytical results, one must be familiar with the materials and painting techniques used through the ages, as well as with the scientific aspects of the techniques, the materials, and their possible reactions; therefore an interdisciplinary work between art historians, conservators, restorers, physicist, and chemists is very important. The results do not help only conservators and restorers but also art historians at understanding better an artist/ workshop/style/influences. Paintings can be carried out on different supports, among which the most used ones are wooden panels, canvas, paper, rarely metal, or glass; sometimes in a museum collection, also a detached wall painting can find its place, being made on plaster. Panels and canvases are first covered by a ground layer (made with chalk or gypsum), and then a thin layer of primer (mostly animal glue or oil) is applied. Sometimes also pigments can be added to the primer for faster modeling. Then, a preparatory drawing is carried out by the principal master, followed by gilding (if chosen), incisions, and pouncing, and at the end the color modeling takes place. To obtain colors, pigments, colorants, and binders are used. There is a wide selection of pigments and colorants; some of them are known since ancient times (as earths and minerals), while many used today were invented only from the second half of the nineteenth century [4–7]. All of them have their proper chemical characteristics that a researcher and a restorer needs to know, and several of them can suffer color changes due to different types of degradation [8, 9]. As a binder, egg yolk, animal glue, or oil are mostly used, sometimes prepared as emulsions. Depending on the binder used, generally also the technique applied gets its name: tempera, oil, acrylic, etc. At the end, the painting is covered by a protective layer or varnish. More about painting techniques can be found in abundant literature [4, 8–11].

23.2

Objectives and Analytical Methods

The principal objective of the study of an artwork is to obtain information on materials that compose it and, when possible, also on painting techniques. Therefore, we are looking into:

644

A. Kriznar et al.

(a) Support – type of wood, panel cloth, metal, etc. (b) Ground and primer (c) Pigments, colorants, binders, and varnish (d) Preparatory drawings, incisions, gilding, modeling, and color layers sequence (e) Results of physical, chemical, and biological damages (f) Previous restoration interventions For this purpose, the traditional method of study is to extract a small sample and study it directly by Raman spectroscopy, or prepared as cross sections under optical microscope, scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDX) and Fourier transformed infrared (FTIR) spectroscopy, or otherwise prepared for the analysis in gas chromatography-mass spectrometry (GC-MS), depending on the information that we are looking for. However, in many cases, the extraction of a sample is not possible; therefore we should use noninvasive analytical techniques. The development of these techniques in recent years has been very fast, and each time we are getting better and more precise results without touching a painting and applying these methods directly in situ, in the exhibition room or in the restoration workshop. Among these techniques are the following: (a) UV-induced visible fluorescence (UVF) photography (b) IR reflectography (IRR) (c) X-ray fluorescence (XRF) and confocal micro X-ray fluorescence (CXRF) spectroscopies (d) Macro X-ray fluorescence (MA-XRF) imaging

23.3

UV-Induced Visible Fluorescence (UVF) Photography

UVF photography is one of the most used techniques in the analysis of artworks, being applied for their examination ever since Wood’s lamps became commercially available around 1925 [12]. This method offers quick information on later interventions, overpaintings, repairs, new varnishes, and false signatures; therefore it can be of great help even for the authentication of a doubtful piece. The ultraviolet (UV) radiation is situated just above the visible light (180–400 nm); nevertheless, for the examination of artworks generally near UV radiation is applied (360 nm). It consists in illuminating an object with UV light (e.g., Wood’s lamps or LEDs-UV), which causes the absorption of the UV radiation by the material. In some cases, the UV radiation provides energy to the material, which results in the emission of fluorescence into visible spectrum. For UV-induced visible fluorescence photography, depending on the UV source, a special filter is, sometimes, needed in order to absorb all the radiation in the visible spectrum; this way, only the desired ultraviolet radiation reaches the studied artwork [13].

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

645

Different materials exhibit different colors and intensities of fluorescence and can vary from light green, yellow, light orange, to brown, depending on the irradiated material. An important fact is that aged materials fluoresce differently than new ones [12, 14, 15]: newer material did not have enough time to establish mutual chemical influences between the pigments and the binders, not even to create molecular compounds that are fluorescent under the UV light. Through this method, we can identify new materials, especially later retouches that appear in the image generally with dark brown colors (Fig. 23.1). Organic materials have a brighter greenish, yellowish glow, so the UV radiation can help localize colorants in comparison to inorganic pigments or old varnishes in comparison to the new ones that still appear dark under the UV light. A very strong bright yellow fluorescence shows zinc white, a relatively new pigment in use since the second half of the nineteenth century [4–7, 16]. Its behavior under UV light reveals later retouches on an older painting or its origin after the second half of the nineteenth century. Dark-bluish color is characteristic of a repair putty used to fill the areas of missing paint, tears, and holes in paintings, while a bluish-white fluorescence indicates lining compound (e.g., if painting layers are lost or very thin and the linen panel can be observed) (Fig. 23.2) [5, 12, 13]. A preliminary identification of pigments is also possible to a certain point, but one must bear in mind that varnish can have a strong fluorescence that, added to the pigment fluorescence, could result in the observation of a different fluorescence color and misguide the interpretation. In consequence, the varnish should be eliminated if UVF is planned to be used for pigment identification, and it must always be complemented with other colorimetric and spectroscopic techniques [16]. However, UVF results might not be always very clear, especially if the analyzed materials contain certain impurities, or if the interventions have been carried out over 80–100 years ago, which gave the latest added material enough time to start fluorescing.

Fig. 23.1 Detail of St. John Evangelist’s coat on The Ascension of the Virgin by Bernardo Martorell (toward 1454), in visible light (a) and in UVF photography (b) revealing many interventions in dark brown color. Museum of Fine Arts, Seville, Spain

646

A. Kriznar et al.

Fig. 23.2 Upper part of Jesus Christ Crucified by Juan de Espinal (1776–1781) in visible light (a) and in UVF before (b) and after (c) the cleaning. Archbishop’s palace, Seville, Spain

23.3.1 Experimental Procedure and Results For good UVF results, the illumination with UV lights should be carried out in the dark, which sometimes is not possible, especially when in the exhibition or restoration room, there is no possibility to dim the windows. In these conditions, working at night could be an option. Different parameters, UV lights, and cameras can be selected for UVF photography, as it is well explained by Cosentino [16]. In our

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

647

studies, we used up four handheld UV lamps, their number depended on the size of the painting. All UV lamps are 6-Watt, model UVGL-55 from Ultra-Violet Products Ltd., which combine long (365 nm) and short (254 nm) waves. The visible fluorescence was registered with a digital camera Nikon D3100 with AF-S DX 18–55 mm and AF-S DX 55–200 mm VR lenses. These small lamps have the advantage of being easy to handle, can be moved in front of the painting surface for observation purposes, or can be fixed on a tripod for UVF photography. It is necessary to use protective glasses to prevent the eyes from the UV radiation. The mentioned procedure was applied in a number of paintings from several periods of time, of different provenance, and composed by variety of materials (supports, pigments, binders, varnish). Two examples were selected for illustration.

Ascension of the Virgin by Bernardo Martorell Bernardo Martorell was a Spanish painter active in the fifteenth century and working in the late international gothic style. Little is known about his life, except that he was one of the leading artists in Catalonia [17]. The panel painting is carried out on wood in tempera technique and represents the Ascension of the Virgin. After her death, the twelve apostles accompany her during the last moments before being called to join her son in Heaven. It is one of the oldest paintings in the collection of the Fine Arts Museum of Seville [17, 18]; therefore it suffered several damages during the centuries, and its image today is the result of past interventions. This is one of the reasons for being chosen for analysis and a future conservation-restoration planification. The first step in the study was the use of UV light to localize retouches and see more clearly the conservation state (Fig. 23.1). Some of the interventions are observed already by the naked eye due to different color hues between the original and the reintegration; however, the UVF photography revealed numerous areas of later retouches, observed as dark brown areas all over the painting surface, especially along the junctions of wooden panels (Fig. 23.1b). It is a classic example of the use of UVF imaging in the analysis of artworks. Jesus Christ Crucified by Juan de Espinal One of the most prolific artists of the eighteenth-century Baroque Seville was Juan de Espinal (1714–1783), whose major opus was carried out for the Archbishop’s palace in the Andalusian capital. According to the documents, between 1776 and 1781 he carried out 15 paintings for the representative stairway of the palace, made on linen canvas in oil technique [19]. Most of them changed their emplacement and some experienced strong humidity that caused heavy damages to color layers and support. One of the most damaged paintings was Jesus Christ Crucified, which is also the only painting in the group with doubtful authorship. When the decision to restore it was taken, also a detailed study of its materials was carried out [20]. Special emphasis was put on UV images (Fig. 23.2), before and after the cleaning of the surface, first to observe the interventions and next to control the cleaning process. On the UVF images obtained before the cleaning (Fig. 23.2b), wide areas of later overpaints and damages can be distinguished, observed as brownish, yellowish, and bluish colors. Clear dark brown lines on the cross, INRI letters, and some other

648

A. Kriznar et al.

brushstrokes of the same color reveal later retouches with an earth pigment, which was confirmed by XRF spectroscopy as umber (Mn, Fe). The bluish fluorescence on the painting borders around the principal figure confirms damages in color layers and their transparency, observed already by the naked eye around the upper borders of the painting (Fig. 23.2a) where linen canvas fluorescence in this color. Similar bluish color emits also binders based on egg tempera, which were clearly used for retouches [12, 15]. The general yellowish-greenish tone of the surface indicates old varnish or general oil binder for original colors. Dark retouches and yellowish varnish were successfully removed, as confirmed by the UVF image after the cleaning (Fig. 23.2c), after which the reconstruction of missing parts, still observed in light blue fluorescence due to the transparent linen canvas, was carried out.

23.4

Infrared Reflectography (IRR)

Since the J. R. J. Van Asperen de Boer thesis, infrared reflectography (IRR) has become a classic image registration method in the technical examination of art [21]. This technique reveals the presence of preparatory drawings made with infrared no-reflective material (such as carbon black) and the artist’s process of execution and changes in the composition (pentimenti) as well [22, 23]. It also provides information that helps to deduce the nature of some pigments present on the artwork’s surface [24, 25] and can give clue about the localization of restoration intervention. Normally, UV-induced visible fluorescence photography is the classic technique used to reveal the presence of the recent organic materials that often is related to restoration interventions. Indeed, it is possible to observe a difference of contrast on the IRR images when the restoration interventions were made using colorant material with different IR reflection properties than the original one. Images under IR radiation have been used for the material study of artwork, mainly in the restoration field, but also as an important tool from the art history perspective. The IRR technique permits users to visualize a painting by registering its response in the infrared spectrum, invisible to the human eye. It has the advantage of being both nondestructive and noninvasive. Infrared radiation extends from the nominal red edge of the visible spectrum at 700 nm to 1 mm. At present, different IRR cameras are commercially available with detectors which are able to record a specific part of the IR spectra depending on the detector technology used (Fig. 23.3). The intensity and the wavelengths of the IR reflected by an artwork depend on the thickness of the paint layers and the kind of materials present (pigments, binding medium, etc.) as well as its concentration.

23.4.1 Experimental Procedure and Results During presented studies, we used an InGaAs camera Xeva-XS-512 model (Fig. 23.4). The Xeva-XS-512 has a resolution of 320256 pixels. In

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

649

Fig. 23.3 Specific part of the IR spectra registered by an InGaAs camera (ISO 20473 IR scheme) [26] Fig. 23.4 The IRR camera Xeva-XS-512 model during in situ measurement at the Fine Arts Museum of Seville

consequence, the reflectograms measured by this camera produced small images, and the equipment must be placed close to the painting (roughly 30 cm in our case) in order to give good results. Depending on the size of the artwork in question, the result can be tens to thousands of images, which must be integrated into a single image. We use the freeware Microsoft Image Composite Editor (Microsoft ICE, version 2.0.3.0) in order to obtain the complete image mosaic. The complexity of obtaining complete IRR images of large format paintings increases proportionally with the number of reflectograms needed [27]. Consequently, precise control of lighting is also necessary to ensure an accurate capture of the painting under the same conditions throughout the reflectographic study [28]. During the present studies, the excitation of the paintings was accomplished by two halogen light reflectors (800 W with intensity regulation control) placed

650

A. Kriznar et al.

on both sides of the artwork. To achieve the most homogeneous illumination, two tripods were placed at the same distance from each side of the painting, both at an angle of 45 with respect to the surface of the painting (Fig. 23.5). Those conditions were applied for every study, in order to homogenize the obtained results.

Fig. 23.5 Positioning of the tungsten lights in order to obtain the most homogeneous illumination during in situ IRR analysis at the Fine Arts Museum of Seville

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

651

The Annunciation Attributed to Alejo Ferna´ndez The Annunciation is an oil panel painting (720495 mm) from San Isidoro del Campo monastery (Seville), exhibited at the Museum of Fine Arts of Seville, dated from 1508 and attributed to Alejo Fernández [29, 30]. Alejo Fernández (c. 1470, Germany (?) – c. 1545, Seville) is considered an important artist for the Renaissance Sevillian painting. He moved to Seville in 1508 and he is best known for The Virgin of the Navigators created as the central panel of an altarpiece for the chapel of the Casa de Contratación in the Alcázar of Seville (between 1531 and 1536) where art historians suggest that Christopher Columbus is represented. The Annunciation painting depicts, on the right, the sitting Virgin receiving the announcement by the archangel Gabriel, placed on the left (Fig. 23.6). They are represented in the middle of an architecture composed of semicircular arches. A column that divides the space in the middle separates the two characters. Through the several apertures of the main building, other architectural structures and a landscape are visible. The IRR study of the Annunciation is composed of 104 IRR images stitched together in one mosaic picture (Fig. 23.7). The high number of images needed was due to the resolution of the IRR camera used but also, to the fact that the painting is protected with a glass. Consequently, in order to take out the reflection of the camera and its support appearing in the protective glass (Fig. 23.8), as much images as possible were taken. Doing so, the stitch process allowed to take out most of the reflection effect in the final mosaic. It was our first attempt to study a painting in such conditions. We can observe that some vertical and horizontal gray strips still remain from the reflection effect. Nonetheless, the IRR mosaic image obtained has sufficient clarity and quality in order to be used for observation purposes. This painting is a perfect study case to show preparatory underdrawings because the artist had previously drawn almost everything represented in the painting as shown by the IRR mosaic image (Fig. 23.7). Moreover, on this image, we can observe the hatching made by the artist in order to localize the shadows on the picture (essentially on the columns and on the ground). The principal artist generally carried out the drawings, so his hand can be identified and the painting related to others. It is, therefore, a very useful tool for art historians when studying style and attribution of an artwork or an artist. This artwork shows also some changes in the composition (pentimenti). The most drastic change is probably the modification of the dimension of the stairs to the left of the archangel and its perspective. One can also compare the visible and IRR images to observe the differences in the folds of the archangel’s tunic. The upper part of the scepter of the archangel has been modified by changing its angle and the shape of its end (an arrow that becomes a ball). Finally, we can see the reduction of the double arch to a single arch for the building at the bottom left. Regarding restoration intervention, IRR mosaic image reveals it most clearly in the lower part of the painting below the two main characters. It is possible to observe different hues of gray, in the IRR image, due to the use for later retouching of different materials than the ones from the original painting.

652

A. Kriznar et al.

Fig. 23.6 The Annunciation attributed to Alejo Fernández (1508), from the Museum of Fine Arts of Seville

San Telmo Attributed to the School of Alejo Ferna´ndez San Telmo is an oil panel painting (1565870 mm) exposed in the Museum of Fine Arts of Seville, from a small church in the neighborhood of San Telmo (Seville) (Fig. 23.9). Hernández related this painting to the early sixteenth century and found similitude with the polyptych from Casa de la Contratación (Seville) attributed to Alejo Fernández [29, 31].

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

653

Fig. 23.7 IRR mosaic image of the Annunciation

San Telmo introduces Blessed Pedro González (sometimes referred to as Saint Telmo), patron of Spanish and Portuguese sailors. Represented as Dominican, he holds in his right hand a blue candle and in his left one, a boat with sailors on board. Behind the rich carpet, a landscape is depicted. The IRR study of San Telmo is composed of 66 IRR images stitched together in one mosaic picture [32]. The IRR mosaic fulfilled our expectations revealing the

654

A. Kriznar et al.

Fig. 23.8 IRR image of the top left corner of the Annunciation. It is possible to see the reflection of the IRR camera on the center

presence of extended underdrawings quite similar to those of the Annunciation. The drawings prepare the elaboration of the principal figure, the shades of the clothing, and the formulation of smaller elements in the composition, such as the architecture in the background or the boat in San Telmo’s left hand. Like in the Annunciation, despite his precise and extended drawing, the artist has changed his composition. Some pentimenti can be seen on the Saint’s head (in the shape of his mouth and the representation of his hair), as well as in the position of his left hand. The most impressive change is in the shape and size of the boat. The original drawing shows in particular a larger crow’s nest surmounted by a triangular flag (Fig. 23.10).

23.4.2 Specific IR Reflectivity of Colorant Materials Infrared false color (IRFC) imaging is the best strategy to differentiate colorant materials using together infrared (NIR) and visible (VIS) images. It is created by digitally editing the VIS and NIR images of a same subject [33, 34]. This composite image linked the visible color of the inks with their IR reflection properties. When compared with a suitable database, IRFC photography is useful as a first attempt to identify the compounds contributing to the ink color [35]. IRFC image is generated by modifying the VIS image inserting the NIR/red channel from the NIR image (Fig. 23.11). The process in Photoshop is as follows: the VIS/blue channel is removed from the VIS image; the VIS/red and VIS/green channels become, respectively, IRFC/green and IRFC/blue channels; the NIR/red channel is imported to the IRFC image and become the IRFC/red channel; and then the IRFC image is converted to RGB mode. When infrared false color imaging is not available, it is still possible to obtain information about pigments and colorants using only IRR images, but it is more

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

655

Fig. 23.9 San Telmo, Anonymous, sixteenth century, from the Museum of Fine Arts of Seville

complicated. To do so, it is important to define a simplified, accurate, and more objective classification for the description of IRR images. The definition presented here has been used to study experimental samples painted with several pigments [25]. This classification contains three categories based on the characteristic IR reflectivity by the colorant materials: – Transparent materials, which reflect a specific part of the IR spectrum – Opaque materials, which do not reflect IR radiation – Shiny materials, which reflect the most part of the IR spectrum reaching the saturation of the detector Our study showed that specifics IR reflectivities are observable with IRR camera regardless of the kind of priming (we test white priming with lead white and red priming with hematite):

656

A. Kriznar et al.

Fig. 23.10 Detail of San Telmo (right) and the IRR mosaic image of the same detail (left), Museum of Fine Arts of Seville

– Lead white, hematite, iron earth yellow, and azurite are transparent. – Malachite is opaque. – Lead-tin yellow and vermilion are shiny. From these observations, we demonstrate that it is possible to distinguish two red pigments (hematite and vermilion (Fig. 23.12), and two yellow pigments (iron earth yellow and lead-tin yellow) (Fig. 23.13), using their specific IR reflectivity property. Nonetheless, the ability to differentiate pigments using their different IR reflectivity properties is more difficult when applied to artworks (also true with IRFC imaging). The reason is that, many times, the artist used a mixture of carbon black or ochre to darken the tone of their colorant material. In consequence, the resulting IR reflection of the mixture is different from the IR reflection of a pure pigment. Because of this, the characteristic IR reflection of the pigment is less consistent. In conclusion, if comparing visible and IRR images of paintings allows a preliminary

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

657

Fig. 23.11 Creating the IRFC image. (a) VIS, (b) NIR, (c) IRFC images of color print Hercules Killing Cacus (posthumous print during the seventeenth century) by the Dutch printmaker Hendrick Goltzius (1558–1617) (412334 mm), from the Edmond de Rothschild Collection from the Louvre Museum [35]. In this example, the pink color from the IRFC image is characteristic of the presence of blue indigotin into the ink

Fig. 23.12 Vermilion (left) and hematite (right), both with white priming, acquired with Xeva-XS-512

assessment of colorant materials, spectroscopic techniques will always be needed in order to determine the nature of the materials presents and to confirm the previous assumptions made based on IRR observations. Imaging techniques are crucial in order to study artworks like paintings or color prints. They provide an overview of the object and facilitate the selection of the punctual analysis area for the spectroscopic techniques such as X-ray fluorescence spectroscopy [23], Fourier transform infrared spectroscopy [36], Raman spectroscopy [37], confocal micro X-ray fluorescence spectroscopy [38], etc.

658

A. Kriznar et al.

Fig. 23.13 Lead-tin yellow (left) and iron earth yellow (right), both with red priming, acquired with Xeva-XS-512

23.5

Portable X-Ray Fluorescence (XRF) Spectroscopy

This noninvasive technique has become one of the most used ones in museums, galleries, or restoration workshops, for a quick in situ (Fig. 23.14) study of materials applied in an artwork [39–41]. A countless number of selected points can be measured, covering all colors, hues, shades, or lights, without damaging the painting [42]; the only limit can be the available time for the analysis or the big dimensions of a painting which can, therefore, not be reached entirely, especially if it cannot be moved from the wall or turned around. This technique is based on the principle of X-ray fluorescence, which occurs basically when an electron in the atom is expelled from its orbit by an X-ray, and another electron from a higher orbit releases part of its energy in the form of an X-ray when it occupies the vacant position. Every atom/chemical element emits its characteristic fluorescence energy, on the basis of which it can be identified [39, 40, 43]. X-ray fluorescence spectroscopy provides elemental analysis of irradiated points, resulting in the identification of the chemical elements in the measured material, particularly pigments and supports in the case of painting analysis. Nevertheless, with this technique, it is difficult to detect elements lighter than Al with the atomic number (Z) lower than 13, when applied in air – the equipment never touches the surface and the measurement is carried out at about 1 cm distance from the analyzed area. Consequently, the XRF technique cannot be used for the identification of organic colorants, binders, varnishes, and pigments composed only by light chemical elements such as ultramarine. This detection limitation can be improved by adding a helium flow in front of the tube/detector system, which can be complicated and quite expensive for in situ measurements, although it has already been achieved [44]. Another disadvantage is that the results are only elemental, since

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

659

Fig. 23.14 Portable XRF spectroscopy in situ measurement in the Museum of Fine Arts, Seville (a), with a closer look at the tube and detector fixed on a metal base (b) and the laser pointing system (c)

XRF does not identify molecular compositions. Because of that, materials with the same principal chemical element but different chemical composition cannot be distinguished, as in the case of copper-based green pigments or lead, cadmium, or chrome pigments [5, 7, 41, 43]. Indeed, the color of the analyzed area can help at the identification of the pigment applied – cadmium yellow or red, for example. However, the problem is also that artists applied mostly mixtures of pigments, which make the interpretation of the XRF spectra even more complicated. Nevertheless, inorganic pigments can be identified on the basis of their principal chemical elements by recognizing chemical elements present in the analyzed area: hematite and earth pigments (Fe), umber (Mn, Fe), cinnabar or vermilion (Hg), azurite (Cu), smalt (Co with Ni, As, Bi impurities), lead pigments (Pb), copper-based green pigments (Cu), zinc white (Zn), titanium white (Ti), chrome pigments (Cr), and cadmium pigments (Cd), to mention the most relevant ones [5, 7, 39]. Inorganic supports, grounds, and priming can as well be identified. Since X-rays penetrate from a few microns to a few millimeters, not only the visible surface is analyzed but also the underlying paint layers. The penetration depth depends on both the energy of the incident X-ray beam (higher for higher energies) and the energy of the emitted X-rays from the material (higher for heavier elements if K-lines are used). With classical XRF spectroscopy, only hypothesis on paint layers can be elaborated; however the relatively recent confocal micro X-ray fluorescence spectroscopy (CXRF) technique is able to identify chemical elements in different layers [38]. In CXRF, a polycapillary lens is attached to the exit of a microfocus

660

A. Kriznar et al.

X-ray tube to reduce the beam diameter to a few tens of μm at the position of the focal point, and another polycapillary lens is placed at the entrance of the X-ray detector. The overlapping of both foci forms the probing micro-volume (or confocal volume), and by scanning through this volume, the composition can be obtained as a function of depth. The depth of analysis is usually limited to less than 1 mm because of the absorption of the X-rays in the matrix. The results of the XRF measurement are observed as spectra, in which all elements found in the analyzed point are represented as peaks that correspond to their K, L, or M lines. For spectra interpretation, one must bear in mind that an artist usually applied a mixture of pigments and that they could be painted in several layers of different thicknesses. If interpreted correctly, not only inorganic pigments and supports can be identified, but also a hypothesis about color layer sequence can be presented. This technique is very suitable also for finding later retouches, identifying pigments that chronologically don't correspond to the age of the painting. Such pigments are, for example, those that were invented after the second half of the nineteenth century on [4, 5, 11].

23.5.1 Experimental Procedure and Results The XRF fluorescence setup mostly applied in our painting studies consists of a W-anode X-ray generator RX38 by EIS with and a silicon drift detector (SDD) with an energy resolution 140 eV at 5.9 keV by Amptek, which also provides the acquisition software of the spectra (Fig. 23.14). An Al filter of 1 mm thickness is coupled to the tube to suppress the W-L peaks overlapping the K lines of Cu and Zn that are important in the identification of pigments. In addition, zirconium (Zr) peaks from the detector collimator usually appear on the spectra, but these do not interfere with the results. The diameter of the radiated spot is 3 mm. The tube and the detector are fixed on a metal base that enables a smooth mechanical forward and backward movement, very important to get to the distance of analysis (Fig. 23.14b). The intersection point of two lasers coupled to the X-ray tube was used to maintain a fixed distance from the X-ray tube to the sample, in such a way that the reproductivity of the geometry tube-sample-detector during the measurements could be guaranteed (Fig. 23.14c). This way, the distance is always controlled without fear to get too close to the painting surface. For all measurements, always the same fixed conditions are used, in order to have the possibility to compare the results of one or several artworks. For our analysis, we use a cathode current of 80 μA and 34 keV of applied high voltage (maximum) and an acquisition live time of 200 s. Analyzed points are often selected together with restorers, according to their necessities, and they cover all the colors and hues of the painting, consulting previously obtained UV and IRR images. As explained before, pigments and supports are identified according to the characteristic energies (keV) of the X-ray peaks in each spectrum, which correspond to specific chemical elements. A quantitative (e.g., with fundamental parameters method) and semi-quantitative (normalized peak area) analysis can be carried out by using the fact that the net peak area of an element is proportional to its concentration.

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

661

Therefore, comparisons between the content of one particular element in different samples of similar composition can be made directly comparing the XRF peaks of that element between different spectra. However, a comparison of the concentrations between different elements can be only possible if a complete calculation (i.e., making the corrections for X-ray production cross sections, self-absorption, etc.) of the concentrations is made. Besides a simple material analysis, this technique can also discover hidden parts of a painting or help us understand the development of an artist through his life and work, as shown by two examples.

The Last Supper by an Anonymous Painter from Luís de Vargas Circle The panel painting made by an anonymous artist around 1570 forms an important part of the collection of sixteenth-century Spanish painting in the Museum of Fine Arts in Seville [17, 19]. According to art historians, it was carried out by a mannerist artist who can be situated in the circle of Luis de Vargas, one of the principal painters in Seville in the sixteenth century. The painting was restored for an exhibition La Huella y la Senda at Las Palmas de Gran Canaria [45], and before the intervention, a material analysis was carried out. The painting represents a well-known subject of the Last Supper (Fig. 23.15), with figures in bright vestments, characteristic for mannerism, in front of a very dark background anticipating what is yet to come. The XRF spectroscopy analysis was selected for pigment analysis and carried out in situ. The results confirmed that pigment selection and their use was very similar to Vargas paintings [46]: lead white (Pb), yellow and red earth pigments (Fe), lead-tin yellow (Pb, Sn), vermilion (Hg), a copper-based green pigment (Cu), smalt (Co, Ni, As, Bi), azurite (Cu), and umber (Mn, Fe). Black color was made with an organic black pigment, impossible to identify with XRF. There is, however, a small difference:

Fig. 23.15 Anonymous: the Last Supper (around 1570). Museum of Fine Arts, Seville. The area of the hidden figure is marked with a square

662

A. Kriznar et al.

Fig. 23.16 XRF spectroscopy results revealed the existence of a figure under the dark background (a), confirmed by the comparison of a spectrum from the background and from the area where the figure was expected (b)

while Vargas used principally smalt for blue color, the anonymous artist applied mostly azurite. High Pb content in every analyzed may come from different sources: (a) priming with lead white; (b) lead-based pigments used in the painting, and (c) lead pigments used as siccative in oil paintings [39, 4, 11]. The results confirmed also the existence of a figure, whose position must have been changed already by the artist himself (Fig. 23.15). By the naked eye, an incised halo can still be observed and some shapes that indicate the existence of a figure; all points out that the apostle in blue was pushed to the right of the table. For this reason, the area in question was measured in several points and compared to other background analyses around it (Fig. 23.16a). XRF results confirmed the existence of carnation under the dark background, made with lead white (Pb), vermilion (Hg), and an earth pigment (Fe), which can be clearly distinguished from other background areas (Fig. 23.16b). Under the black color, also a gilded halo (Au) probably on bole (Fe) was confirmed, as well as apostle’s blue tunic, made with a mixture of azurite (Cu) and lead white (Pb).

Gonzalo Bilbao and His Opus Gonzalo Bilbao (1860–1938) is considered one of the most outstanding artistic personalities of Spanish painting at the turn of the nineteenth to the twentieth centuries [17, 19]. An important exhibition dedicated to him was organized by the Fine Arts Museum of Seville, with a selection of works created throughout his entire life and coming from different collections [47]. For this purpose, many of them were restored, and 14 were selected for analysis by XRF spectroscopy, in order to know the materials and painting techniques of the artist. Another objective was also to determine if there were any changes in the choice of materials, bearing in mind that Bilbao was living in the time when many new pigments were invented, and artists have a wide possibility of experimenting with them [4, 5, 10, 11]. Therefore, works

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

663

from his entire artistic life were selected for analysis; the earliest painting being Portrait of Elena Sánchez from 1890 and the latest, Portrait of Francisco Rodriguez de Marín from 1934. The results clearly show important changes and developments in his work, observing the use of different materials in the priming and in his pigment choice [47, 48]. In his early works, Bilbao used for the preparation of his canvases only lead white (Pb), a common material for preparation since the beginning of canvas painting in the fifteenth century [49]. Approaching the end of the nineteenth century, he started to introduce the new pigment zinc white (Zn) and gradually replacing lead white with it. Finally, in his last works he only applied lithopone (Zn, Ba) (Fig. 23.17a). Results obtained on pigments are even more complex: Bilbao was introducing new pigments almost as they were appearing on the market, while at the same time, he gradually stopped using traditional ones from his early period [46, 47]. In his early works before 1900, almost exclusive use of traditional pigments can be observed, as lead white (Pb), yellow and red earth pigments (Fe), umber (Mn, Fe), vermilion (Hg), bone black (Ca), and a copper-based green pigment (Cu). There are already some modern pigments present as zinc white (Zn), chrome green (Cr), cadmium yellow (Cd), or cobalt blue (Co), but their use is limited. Between 1900 and 1910, these modern pigments start to gain importance, and Bilbao introduces new pigments as strontium yellow (Cr, Sr) and cobalt violet (Co, As) and also experiments with some other pigments which he stops using after 1910, like the intense copper and arsenic green that could be Scheele’s (Cu, As) or Schweinfurt/Emerald (Cu, As) green. After 1910 his palette gets even more complex and it is difficult to identify the pigments only by XRF spectroscopy, being several characterized by same chemical elements. Nevertheless, it is possible to confirm the introduction of cadmium orange (Cd) and cerulean blue (Co, Sn). There is an important change in the basic palette: the use of lead white and zinc white becomes practically the same, showing that the painter used both in the same quantity. In his late artistic period around and after 1920, zinc white became his principal white pigment, introducing also lithophone (Ba, Zn), while lead white practically disappeared. Earth

Fig. 23.17 Comparison of XRF spectra obtained on Bilbao’s paintings from different periods, showing changes in priming layers (a) and in the use of pigments for carnations (b)

664

A. Kriznar et al.

pigments and vermilion were also slowly substituted with cadmium yellow (Cd) and red (Cd, Se – placed on the market only after 1910), as can be well observed in comparison carnations from different periods of his work. In his early career, he modeled carnations with lead white and vermilion, while at the end he used mostly zinc white or lithophone and cadmium red (Fig. 23.17b). He continued experimenting with new pigments, for instance, phthalocyanine green (Cl, Cu). However, the use of those new pigments cannot be confirmed by XRF with certainty due to the coincidence of the characteristic chemical elements or for being organic colorants.

23.6

Macro X-Ray Fluorescence (MA-XRF) Imaging

The application of X-ray fluorescence (XRF) could be rather complex in the case of paintings. Paintings are large dimension samples; they exhibit a complex layered structure and a heterogeneous multi-elemental composition. Application of XRF as a single spot analysis could provide misleading information. In 2008, the first XRF scanning of a medium-sized painting at a synchrotron facility allowed a simultaneous mapping of chemical elements in the pictorial support [50]. Over the time, this experimental approach has been optimized and mobile instruments based on a X-ray tube irradiating sources were developed with the aim of performing the macroscopic XRF (MA-XRF) scanning of easel paintings [51–54]. MA-XRF is performed by scanning samples with an X-ray beam presenting a dimension of few hundred microns. The primary radiation penetrates the painting and induces chemical elements in the pictorial layers to emit the characteristic X-ray fluorescence. The X-ray fluorescence is properly detected, and it is used to obtain the elemental distribution images in the pictorial support with a lateral resolution defined mainly by the beam-size and the scanning step-size. The two-dimensional elemental images contain integrated information on the composition of pigments along the artwork thickness. Benefits of using MA-XRF (even in combination with other imaging techniques) have been verified in different analytical case study [55]. The technical investigation of paintings carried out using the MA-XRF allowed to better elucidate the nature of pigment materials, the painting technique, as well as to put in evidence the presence of degradation processes and to investigate the conservation state of the artworks. In some cases, the analysis allowed to approach question of attribution and authenticity. Finally, thanks to the multispectral and penetrating capabilities of the technique allowed to gain information on the pigments beyond the visible surface and to put in evidence the existence of hidden details that are considered of great importance for understanding the creative process of an artist in his pictorial composition.

23.6.1 Experimental Procedure and Results The elemental distribution images of the painting discussed in this work were obtained by using the LANDIS-X system, a novel mobile X-ray scanner based on

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

665

a real-time technology developed at the LANDIS laboratory of the ISPC-CNR and LNS-INFN in Catania (Italy). The system integrates macro-XRF (MA-XRF) and micro-XRF (μ-XRF) for performing in situ the two-dimensional elemental imaging of artworks with high scanning speed and high lateral resolution [52]. LANDIS-X is equipped with a spectrometric head composed of a Rh-anode microfocus X-ray source, coupled to a polycapillary lens in the excitation channel and two SDD detectors positioned in a 45-90-45 degree geometry and operated in parallel. The spectrometric head is moved during the scanning by means of a threeaxis motorized stage presenting a travel range of 1107020cm3. The XY stages cover the flat area of the painting under investigation while the Z-axis is dynamically guided with a laser sensor for keeping the spectrometric head at a fixed measurement distance during the scanning. A rotational scanning mode is also available for the elemental imaging of artworks, presenting a three-dimensional geometry. The combination of the rotational scanning of the sample positioned on an ancillary rotational stage with the translation of the spectrometric head along the vertical direction enables the unrolling of the curved sample surface providing the elemental distribution images over a 360 degrees view. MA-XRF analysis is carried out by positioning the samples out of the polycapillary focus. The typical measurement distance is 10 mm where the primary beam presents a spot size of few hundreds microns and a high intensity. We estimate an input counting rate (ICR) on the two detectors > 2.5Mcps with the X-ray source operated at 50kV and 0.6mA. The use of state-of-the-art digital X-ray processor based on a time-list-event-mode (TLIST) acquisition allows us to work at such count rates with a reduced dead time. The full area of 11070 cm2 is covered in less than 2 h at 100 mm/sec. The μ-XRF analysis is performed by positioning the sample at the focus (i.e., at 15 mm) where the beam presents a 50 μm spot-size. The real-time capabilities of the scanner allow us to achieve imaging details down to the micrometric scale within a macroscopic context. High definition elemental distribution images over 15 megapixels are obtained in about 20 h. LANDIS-X is fully controlled by a custom-developed control unit. X-ray pixel spectra are processed in real time by applying a least square fitting procedure; the deconvoluted elemental distribution images are elaborated during the scanning and are available in live mode to the users (Fig. 23.18).

Virgin with a Child by a Lorenzo di Credi At the beginning of the sixteenth century, the workshop of Lorenzo di Credi [56] was one of the most active centers in the production of devotional paintings in Italy. Due to the success of his artistic work, apprentices in his workshop and followers started to imitate his style [57]. In addition, the request of his paintings in the market of art at the end of the nineteenth century encouraged the production of a large number of fakes. MA-XRF applied to paintings of doubtful attribution can be used to verify the compatibility of pigments with the ones in the palette of the artist or to detect the presence of anachronistic pigments.

666

A. Kriznar et al.

Fig. 23.18 The mobile LANDIS-X scanner during a real-time MA-XRF imaging of an historical painting

A painting attributed to Lorenzo di Credi entitled Virgin with a Child (oil on panel, 1503, 4057 cm2) (Fig. 23.19) was investigated by MA-XRF in situ at the Museum of Castello Ursino in Catania (Italy). The preparation of the panel is made with a calcium-based material and lead white as evidenced by the calcium and lead distribution maps that characterize the whole pictorial surface of the painting (Fig. 23.19). Compounds containing calcium are also visible in the damaged areas of the painting. They were used as filling materials for the lacunas during the restoration works performed over the time on the painting. The artist used vermilion to paint the dress of the Virgin and, in combination with lead white, for the pink tones of the faces of both figures. Vermilion is also combined with iron-based pigments in the hair of the Virgin and of the child. Iron-based earth pigments characterize the tree depicted on the right side of the panel and they have been used to obtain chiaroscuro effects. Titanium is spatially correlated with iron suggesting its presence as an impurity of the earth’s pigments. The distribution of all these elements can be well observed in Fig. 23.20. The distribution of cobalt (Co) and potassium (K) (Fig. 23.21) identifies the use of smalt for the drape on Virgin’s dress, a synthetic pigment used mostly between the fifteenth and the seventeenth centuries, obtained by the pulverization of a potassium enriched glass to which cobalt oxide is added [5–9]. In addition, it includes Ni, As, and Bi, typical trace elements of the cobalt mineral used as primary raw material in the small manufacturing. Their distribution can be as well observed in Fig. 23.21. Traces of K also

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

667

Fig. 23.19 Madonna with Child by Lorenzo di Credi (ca. 1503) with the elemental distribution maps of calcium (Ca) and lead (Pb) obtained by MA-XRF imaging during the scanning

Fig. 23.20 The elemental distribution maps of mercury (Hg), iron (Fe), and titanium (Ti) on Madonna with Child

appear in the regions of the painting where no Co is present, and this allows us to assume the use of a lake of organic nature (probably mixed with alum as binder) to which the MA-XRF technique is not sensitive. The pigments based on copper are compatible with the use of different historical pigments. Azurite can be attributed to the blue pigment in the paintings, while the attribution of the green tones is more difficult by the use of the elemental map of copper only, since verdigris, copper resinate, or malachite are all possible due to their extensive use in paintings over the fifteenth and sixteenth century [5–8, 10].

668

A. Kriznar et al.

Fig. 23.21 The elemental maps of cobalt (Co), potassium (K), and arsenic (As) (top); the maps of nickel (Ni), bismuth (Bi), and copper (Cu) (bottom)

MA-XRF has not detected the use of anachronistic pigments (e.g., titanium white or zinc); therefore, modern retouches or restoration of the panel could be excluded based on the MA-XRF analytical result.

23.7

Conclusions

Noninvasive analyses with their huge development in the last decades have become one of the principal parts in the study of artworks in museums, galleries, and restoration workshops, offering in situ measurements and fast information on

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

669

material and artistic techniques without touching the object. The most important methods were presented in this chapter, sharing some outstanding results. Through UVF and IRR, hidden parts of a painting can be discovered, as later retouches, pentimenti, or even some pigment identification. The material characterization can be well achieved through XRF, μ-XRF spectroscopies, and MA-XRF imaging, which even reveals the distribution of chemical elements on the surface. Many museums can already afford their UVF, IRR, and XRF equipment, especially purchasing smaller handheld instruments; however for better and more complex results, a collaboration with research groups that work with other techniques such as MA-XRF, Raman, FTIR, and CXRF spectroscopies is necessary. Despite more and more precise instrumentation, the interpretation of the results can still be complicated or even doubtful. For this reason, the interdisciplinary collaboration between humanists and scientists is of major importance, and both sides should work side by side. Acknowledgments We wish to thank Mª del Valme Muñoz Rubio, director of the Fine Arts Museum of Seville; Fuensanta de la Paz, Mercedes Vega, and Alfonso Blanco, from the restoration department of the Museum; Ana Isabel Gamero González, conservator of the Cultural Heritage of the Archdiocese of Seville; and Carolina López, Juanjo Mañas, Toni Ruiz, and Gerard Ester, from Infaimon for their support. This work was partially supported by the Junta de Andalucía Projects of Excellence 2005/HUM-493 and P09-HUM-4544], Spanish Ministry of Science and Innovation (Juan de la Cierva research contract), and the Fondation des Sciences du Patrimoine/LabEx PATRIMA ANR-10-LABX-0094-01, « CLARO » project (EUR-17-EURE-0021). The author wishes to acknowledge professional support of the Interdisciplinary Thematic Platform from CSIC Open Heritage: Research and Society (PTI-PAIS).

References 1. Ankersmit B, Stappers MHL (2018) Managing indoor climate risks in Museums, cultural heritage science. Springer International Publishing, Basel 2. Berger GA (2000) Conservation of paintings, research and innovations. Archetype Publications, London 3. Mitchell R, Clifford J (eds) (2018) Biodeterioration and preservation in art, archeology and architecture. Archetype Publications, London 4. Kühn H, Roosen-Runge H, Straub RE, Koller M (eds) (1990) Reclams Handbuch der künstlerischen Techniken, vol 1. (Philipp Reclam jun., Stuttgart) 5. Fitzhugh EW, Feller RL, Roy A, Berrie B (eds) (2012) Artists’ pigments: a handbook of their history and characterisation. Archetype Publication, London 6. Schweppe H (1993) Handbuch der Naturfarbstoffe. Nikol Verlagsgesellschaft mbH&Co.KG, Hamburg 7. Eastaugh N, Walsh V, Chaplin T, Siddall TR (2008) Pigment Compendium, a dictionary and optical microscopy of historical pigments. Elsevier, New York/London 8. van Eikema Hommes M (2004) Changing pictures, discoloration in 15th-17th century oil paintings. Archetype Publications, London 9. Cocciato A, Moens L, Vandenabeele P (2017) On the stability of mediaeval inorganic pigments: a literature review of the effect of climate, material selection, biological activity, analysis and conservation treatments. Herit Sci 5(12):2–25. https://doi.org/10.1186/s40494-017-0125-6 10. Laurie AP (1967) The Painter’s methods and materials, traditional techniques and materials for practicing artists: oil, watercolor, tempere. Dover, New York

670

A. Kriznar et al.

11. Brachert T (2001) Lexikon historischer Maltechniken, Quellen-Handwerk-TechnologieAlchemie. Callwey, München 12. de la Rie ER (1982) Source: fluorescence of paint and varnish layers. Stud Conserv, Part I: 27(1):1–7, 10.2307/1505977; Part II: 27(2):65–69 (1982), 10.2307/1505989; Part III: 27(3): 102–108 (1982). https://doi.org/10.2307/1506145 13. Laclavetine K (2015) Non-destructive techniques for in situ studies of easel paintings: development and application of a confocal X-ray micro-fluorescence system. PhD thesis, University of Seville, Sevilla 14. Matteini M, Moles A (1994) Scienza e restauro. Metodi d´indagine. Nardini, Firenze 15. Aldrovandi A, Picollo M (2003) Metodi di documentazione e indagini non invasive sui dipinti. Il Prato, Padua 16. Cosentino A (2015) Practical notes on ultraviolet technical photography for art examination. Conservar Patrimonio 21:53–62. https://doi.org/10.1063/1.3174431 17. Izquierdo R, Muñoz V (1990) Museo de Bellas Artes, Inventario de pintura. Grafiberica, Sevilla 18. Kriznar A, Muñoz MV, Respaldiza MA, Vega M (2012) Materials applied in Bernardo Martorell’s painting analysed by portable XRF. ArchéoSciences 36:37–45 19. Valdivieso E, Serrera JM (1979) Catálogo de las pinturas del palacio arzobispal de Sevilla. Editorial Sever-Cuesta, Sevilla (España) 20. Križnar A, Laclavetine K, Gamero González AI, Respaldiza MA (2015) Análisis no destructivo del “Cristo Crucificado” de Juan de Espinal. In: Moreno Oliva MM, Rogiero Candelera MA, López Navarrete JT, Hernández Jolín V (eds) Estudio y conservación del patrimonio cultural, Actas TechnoHeritage 2015. University of Málaga, Malaga, pp 102–106 21. van Asperen de Boer JRJ (1968) Infrared reflectography: a method for the examination of paintings. Appl Optics 7(9):1711–1714. https://doi.org/10.1364/ao.7.001711 22. Laclavetine K (2018) Técnicas espectroscópicas y multiespectrales aplicadas al estudio no destructivo in situ de un conjunto de obras atribuidas al pintor Alonso Vázquez. In: Innovación y nuevas tecnologías conservación y restauración. Ediciones Trea, Bilbao, pp 17–56 23. Kriznar A, Laclavetine K, Muñoz V, Respaldiza MÁ, Vega M (2016) Non-destructive analysis of pigments in a triptych by Marten de Vos. Spectrosc Lett 49(1):30–36. https://doi.org/10.1080/ 00387010.2015.1061018 24. Casini A, Lotti F, Picollo M, Stefani L, Buzzegoli E (1999) Image spectroscopy mapping technique for noninvasive analysis of paintings. Stud Conserv 44(1):39–48. https://doi.org/10. 1179/sic.1999.44.1.39 25. Laclavetine K, Hernández E, Arquillo J, Arroyo E, Respaldiza MA (2013) Análisis de estándares de pintura según recetas del siglo XVI mediante reflectografía infrarroja (IRR) (4th Latin-American Symposium on Physical and Chemical Methods in Archaeology, Art and Cultural Heritage Conservation (LASMAC2013)). Bogota, Colombia, p 6 26. Miller JL (1994) Principles of infrared technology: a practical guide to the state of the art. Springer, New York 27. Hernández E, Laclavetine K, Alvarado R, Kriznar A, Muñoz MDV, Vega M, Arroyo E, Ruvalcaba Sil JL, Respaldiza MA (2014) Infrared reflectography (IRR) for large formats: comparison of various capture and mosaic assembly methodologies for large format easel paintings. In: 11th international conference on non-destructive investigations and microanalysis for the diagnostics and conservation of cultural and environmental heritage, Madrid, Spain, p 8 28. Consolandi L, Bertani D (2007) A prototype for high resolution infrared reflectography of paintings. Infrared Phys Technol 49(3):239–242. https://doi.org/10.1016/j.infrared.2006.06.031 29. Valdivieso González E (1993) La pintura en el Museo de Bellas Artes de Sevilla. Ediciones Galve, Sevilla 30. Gómez Sánchez JA (2016) Alejo Fernández y la pintura sevillana del primer tercio del siglo XVI. PhD thesis, University of Seville, Sevilla 31. Hernández Díaz J (1967) Museo Provincial de Bellas Artes de Sevilla Guías de los Museos de España XXX. Ministerio de Educación y Ciencia, Sevilla

23

Noninvasive Imaging and Spectroscopic Techniques Applied In Situ in Museums

671

32. Kriznar A, Laclavetine K, Muñoz MV, Respaldiza MA, Vega M (2015) Non-destructive techniques for in situ studies of a 16th century panel painting. Opuscula Musealia 23:113– 120. https://doi.org/10.4467/20843852.OM.15.010.5388 33. Cosentino A (2016) Infrared technical photography for art examination. e-Preservation Science 13:1–6 34. Moon T, Schilling MR, Thirkettle S (1992) A note on the use of false-color infrared photography in conservation. Stud Conserv 37(1):42–52. https://doi.org/10.2307/1506436 35. Laclavetine K, Boust C, Clivet L, Le Hô AS, Laval E, Mathis R, Menu M, Pagliano E, Salmon X, Selbach V, Vrand C, Lepape S (2019) Non-invasive study of 16th century Northern European chiaroscuro woodcuts: first insights. Microchem J 144:419–430. https://doi.org/10. 1016/j.microc.2018.10.003 36. Buti D, Rosi F, Brunetti BG, Miliani C (2013) In-situ identification of copper-based green pigments on paintings and manuscripts by reflection FTIR. Anal Bioanal Chem 405(8):2699– 2711. https://doi.org/10.1007/s00216-013-6707-6 37. García-Bucio MA, Casanova-González E, Ruvalcaba-Sil JL (2014) Raman spectroscopy for the study of XVI-XVII Centuries Colonial Paintings. MRS Online Proceedings Library 1618:141– 151. https://doi.org/10.1557/opl.2014.463 38. Laclavetine K, Ager FJ, Arquillo J, Respaldiza MÁ, Scrivano S (2016) Characterization of the new mobile confocal micro X-ray fluorescence (CXRF) system for in situ non-destructive cultural heritage analysis at the CNA: μXRF-CONCHA. Microchem J 125:62–68. https://doi. org/10.1016/j.microc.2015.11.013 39. Seccaroni C, Moioli P (2004) Fluorescenza X: Prontuario per l’analisi XRF portatile applicata a superfici policrome. Nardini editore, Firenze 40. Neelmejer C, Brissaud I, Calligaro T, Demortier G, Hautojarvi A, Mader M, Martinot L, Schreiner M, Tuurnala T, Weber G (2000) Paintings – a challenge for XRF and PIXE analysis. X-Ray Spectrometry 29:101–110. https://doi.org/10.1002/(SICI)1097-4539(200001/02)29: 13.0.CO;2-A 41. Deming Glinsman L (2005) The practical application of air-path X-ray fluorescence spectrometry in the analysis of museum objects. Stud Conserv 50:3–17. https://doi.org/10.1179/sic. 2005.50 42. Mantler M, Klikovits J (2004) Analysis of art objects and other delicate samples: is XRF really nondestructive? Powder Diffr 19(1):16–19. https://doi.org/10.1154/1.1649962 43. Artioli G (2010) Scientific methods and Cultural Heritage: an introduction to the application of materials science to archaeometry and conservation science. Oxford University Press, Oxford/New York 44. de Viguerie L, Walter P, Laval E, Mottin B, Solé VA (2010) Revealing the sfumato technique by Leonardo da Vinci by X-ray fluorescence spectroscopy. Angew Chem Int Ed 49(35):6125– 6128. https://doi.org/10.1002/anie.201001116 45. Lavandera López J, Cova del Pino JF (2004) La Huella y la Senda, exhibition catalogue, Catedral de Santa Ana, Las Palmas de Gran Canaria, 30th January – 30th May of 2004. Gobierno de Canarias, Viceconsejería de Cultura y Deportes, Las Palmas de Gran Canaria, in Spanish 46. Kriznar A, Muñoz MV, Respaldiza MA, Vega M (2011) La Cena Sacramental del pintor anónimo perteneciente al círculo de Luís de Vargas (s. XVI) analizado por la técnica no-destructiva de XRF. In: López Montes AM, Collado Montero F, Medina Flórez V, Espejo Arias T, García Bueno A (eds) XVIII Congreso Internacional Conservación y Restauración de Bienes Culturales, Granada 9–11 noviember 2011, Libro de actas. Universidad de Granada, Granada, pp 211–214 47. Márquez V (ed) (2011) Gonzalo Bilbao. Fondos del Museo de Bellas Artes de Sevilla, November 2011 – March 2012, Exhibition catalogue. Junta de Andalucía, Sevilla, in spanish 48. Kriznar A, Valme Muñoz M, Respaldiza MA, Vega M (2013) Pigments and painting preparations of Gonzalo Bilbao analysed by non-destructive XRF technique. In: Rogerio-Candelera

672

A. Kriznar et al.

MA, Lazzari M, Cano E (eds) Science and technology for the conservation of cultural heritage. CRC Press/ Balkema, Taylor & Francis Group, London, pp 185–188 49. Townsend JH, Doherty T, Heydenreich G, Ridge J (eds) (2008) Preparation for painting, the artist’s choice and its consequences. Archetype Publications, London 50. Dik J, Janssens K, Van der Snickt G, Van der Loeff L, Rickers K, Cotte M (2008) Visualization of a lost painting by Vincent van Gogh using synchrotron radiation based X-ray fluorescence elemental mapping. Anal Chem 80(16):6436–6442. https://doi.org/10.1021/ac800965g 51. Alfeld M, Vaz Pedroso J, van Eikema Hommes M, Van der Snickt G, Tauber G, Blaas J, Haschke M, Erler K, Dik J, Janssens K (2013) A mobile instrument for in situ scanning macroXRF investigation of historical paintings. J Anal At Spectrom 28:760–767. https://doi.org/10. 1039/C3JA30341A 52. Romano FP, Caliri C, Nicotra P, Di Martino S, Pappalardo L, Rizzo F, Santos HC (2017) Realtime elemental imaging of large dimension paintings with a novel mobile macro X-ray fluorescence (MA-XRF) scanning technique. J Anal At Spectrom 32(4):773–781. https://doi. org/10.1039/C6JA00439C 53. Alberti R, Frizzi T, Bombelli L, Gironda M, Aresi N, Rosi F, Miliani C, Tranquilli G, Talarico F, Cartechini L (2017) CRONO: a fast and reconfigurable macro X-ray fluorescence scanner for in-situ investigations of polychrome surfaces. X-Ray Spectrom 46(5):297–302. https://doi.org/ 10.1002/xrs.2741 54. Ravaud E, Pichon L, Laval E, Gonzalez V, Eveno M, Calligaro T (2016) Development of a versatile XRF scanner for the elemental imaging of paintworks. Appl Phys A 122:17–24. https://doi.org/10.1007/s00339-015-9522-4 55. Alfeld M, de Viguerie L (2017) Recent developments in spectroscopic imaging techniques for historical paintings – a review. Spectrochimica Acta Part B 136:81–105. https://doi.org/10. 1016/j.sab.2017.08.003 56. Brewer R (1970) A study of Lorenzo di Credi. Tip. Giuntina, Firenze 57. Zeri F (ed) (1986) La pittura in Italia. Il Quattrocento. Banca nazionale dell’agricoltura, Venice

Archaeometallurgy: A Discipline Between Past and Future

24

Claudio Giardino

Contents 24.1 24.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Brief History of Archaeometallurgy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.1 The Precursors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2 The Dawn of Archaeometallurgy as a Science: Between the Eighteenth and Nineteenth Centuries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.3 The Second Half of the Nineteenth Century . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.4 The New Approaches: Twentieth Century . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.5 The Recent Developments in Archaeometallurgical Research: An Outlook for the Future (End of the Twentieth Century to Beginnings of the Twenty-First Century) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3 Analytical Techniques and Methodologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.1 General Requirement for Archaeometallurgical Investigations . . . . . . . . . . . . . . . 24.3.2 Nondestructive and Noninvasive Investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.3 Minimally Invasive Investigations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

674 675 675 675 679 679

683 685 685 688 692 696

Abstract

1. A Brief History The determining role of metals in the development of civilization was already understood in antiquity. Since the end of the eighteenth century ancient metal objects started to be scientifically characterized. In the first half of the nineteenth century, the Positivists thought that the chemical investigations of ancient metals could solve historical and archaeological problems. Ancient metals became the object of relevant, ambitious, and extensive analytical investigations from the mid-1900s. At the same time, scholars such as Gordon Childe, Leroi-Gourhan, Stanley Smith, Tylecote, and Chernykh created a theoretical framework for

C. Giardino (*) University of Salento, Lecce, Italy © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_24

673

674

C. Giardino

archaeometallurgical studies. By the end of the twentieth century, relevant projects were published and new research fields explored. 2. Analytical Techniques Archaeometallurgy is a multidisciplinary research field, where the analyses are carried out on archaeological and artistic artifacts to provide an answer to specific problems. The study of ancient metals is a preferential research field for nondestructive and noninvasive investigations, for example, X-rays, γ rays, ultrasonics, acoustic emission, and XRF. Anyhow, modern analytical techniques allow taking samples of very small dimensions: we can speak of minimally invasive investigations, for example, spectrometry (MS, NAA, AES) and metallography (optical and SEM). ICP-MS plays relevant roles in provenance studies by detecting the Pb isotopic composition.

24.1

Introduction

Archaeometallurgy is a typically multidisciplinary discipline that studies metals and their use in past societies. Metal artifacts and objects related to their production – such as crucibles, casting molds, etc. – are examined not only from a compositional point of view but also in the way in which they were made (the so-called operational chain) and in their function in different human societies, therefore, from a historical and anthropological perspective. Metals have always had significant social and economic importance: archaeometallurgy, therefore, examines how metal technologies were integrated into the social and economic life of the communities that made and used them. The investigations analyze various aspects related to metallurgy, such as mineral prospecting and the extraction of minerals from the subsoil, therefore the ancient mining techniques; methods of extracting metal from ore; how the artifacts were produced, the casting, and smithing techniques; the use of objects in everyday life, wconnected with use-wear studies; the traffic and trade routes that allowed the transport of minerals, an assemblage of metal as ingots and finished objects from one region to another, even at great distances; and the analytical techniques used in the investigations. Most metal objects are subject to corrosion: a branch of archaeometallurgy studies the characteristics of decay and what strategies allowed it to be prevented. Since metals have a universal diffusion, their study must take into account the different development of metallurgy in different historical and geographical contexts, not only from the technological point of view but also from cultural and social aspects. Like all scientific studies, archaeometallurgy is a dynamic discipline and is subjected to continuous evolution, not only in the investigative techniques but also in the more strictly theoretical and heuristic aspects. New data and new investigation techniques can therefore revolutionize paradigms previously considered certain.

24

Archaeometallurgy: A Discipline Between Past and Future

24.2

675

A Brief History of Archaeometallurgy

24.2.1 The Precursors The importance of metals from an economic and social point of view was clear since antiquity: the Greek poet Hesiod (late eighth to early seventh century BC), in his work Works and Days, made a first attempt at reconstructing the stages of human development based on metals – the ages of Gold, Silver, Bronze, Heroes, and Iron. The encyclopedic treatise of Pliny the Elder (23–79 AD) Naturalis Historia occupies particular relevance in the studies on ancient metallurgy. Pliny extensively discusses metals, metallurgy, and minerals in books XXIII and XXIV, providing an articulated vision on the knowledge of these topics possessed by the Roman world. In the Middle Ages, the main treatise on the metallurgy of the time was the De Diversis Artibus from the German monk Theophilus (pseudonym of Roger of Helmarshausen: c. 1070–1125). In the Orient, the Iranian scholar Abū Rayḥān Muḥammad ibn Aḥmad Al-Bīrūnī (973–after 1050) and the Arab geographer Muḥammad Abū’l-Qāsim Ibn Ḥawqal (943–969) described the work in the mines in their books (among them Ṣūrat al- ‘Arḍ, “The face of the Earth” of 977). In the Renaissance, important treatises on contemporary metallurgy were written, such as De los metales (1530) by Bartolomé Pérez de Veiga, De la pirotechnia (1540) by Vannoccio Biringuccio, and De re metallica (1556) by Georgius Agricola (George Bauer: 1494–1555). This last work in 12 books represents the most exhaustive treatise on metallurgy and mining of the sixteenth century and is considered the first scientific account on these topics, still widely used in archaeometallurgical studies also for its vast iconography apparatus. Sebastian Münster’s Cosmographia universalis (1544) is also worth mentioning, where the German cosmographer accurately illustrates, among other things, the techniques related to the world of mines (Fig. 24.1), and Benvenuto Cellini’s Two Treatises (1568), in which the Italian artist illustrates goldsmithing and bronze working techniques (Fig. 24.2).

24.2.2 The Dawn of Archaeometallurgy as a Science: Between the Eighteenth and Nineteenth Centuries Since the beginning of archaeological studies, metal objects have attracted the scholars’ interest, in the belief that metals were decisive in reconstructing the societies and the cultures of the past. The earliest, tentative steps toward a systematic investigation of ancient metals took place in Europe in the seventeenth and eighteenth centuries, during the Enlightenment era. At the beginning of the nineteenth century, Christian Jürgensen Thomsen (1788– 1865) attributed to metals the role of chronological markers for the change of ages in his system of the three ages, as previously done by Hesiod in antiquity. He created his classification for the Danish National Museum of Archaeological Collections, according to whether the artifacts were made of stone, bronze, or iron [1] (Fig. 24.3).

676

C. Giardino

Fig. 24.1 Metal ore beneficiation, from Sebastian Münster, Cosmographia universalis, liber III

The first archaeometric analyses of ancient metals date from the late eighteenth century. These earliest investigations were solely aimed at determining the main alloy components, without paying attention to the minor elements, since it was not yet possible to carry out complete quantitative chemical analyses. Furthermore, there was still no precise chronological attribution of the examined findings: the Thomsen’s

24

Archaeometallurgy: A Discipline Between Past and Future

677

Fig. 24.2 Benvenuto Cellini’s “Due Trattati,” frontispiece

system was only developed in the 1830s. The first investigations had heterogeneous material objects for both typology and chronology, such as coins, axes, daggers, swords, and rings [2–7]. At the beginning of the nineteenth century, investigations of ancient finds increased, with more precise measurements of the alloy components, thanks to the development of more advanced analytical techniques. These

678

C. Giardino

Fig. 24.3 Christian Jürgensen Thomsen (portrait by J.V. Gertner, 1818–1871). (https://da.wikipedia.org/ wiki/Christian_Jürgensen_ Thomsen#/media/Fil: Christian_Jürgensen_ Thomsen.jpg)

investigations addressed not only bronze finds and their patinas but also precious metals, iron, and slag [8–10]. Ancient mines in the Iberian Peninsula and Greece were also studied [11–13]. In the first half of the nineteenth century, positivism began to develop in Europe a philosophical paradigm that exalted science as the main – if not the only – source of knowledge. This thought, developed by the French philosopher Auguste Comte (1798–1857), characterized the second half of the century. In archaeometallurgical studies, positivism inserted itself into the studies of antiquities that were still predominant since the beginning of the nineteenth century. Now chemical investigations of ancient metals aimed at “scientifically” integrating technological data, to solve historical and archaeological issues [14, 15]. The new information thus obtained would have made it possible to better understand the artifacts and to know the origin of the raw materials, a matter that has always been a relevant problem for archaeologists. One of the earliest large-scale archaeometallurgical investigation programs – 120 analyses – was conducted and published by Fr. Göbel in the first half of the nineteenth century [16]. The title of the study, “On the influence of chemistry on the identification of prehistoric people,” explicitly suggested the potential of scientific methods to clarify archaeological problems.

24

Archaeometallurgy: A Discipline Between Past and Future

679

Geographic explorations and related ethnological investigations – linked to the phenomenon of European colonialism in Africa and Asia – also contributed to developing new knowledge on the metallurgy of non-European populations. The British metallurgist John Percy (1817–1889), based on reports of Captain Dixon in the 1830s [17], examined the contemporary metallurgical techniques of India and South Asia, including the beneficiation techniques and the performance of Indian smelting furnaces, attempting a first historical reconstruction of smelting techniques [18].

24.2.3 The Second Half of the Nineteenth Century The study of ancient mines was greatly expanded thanks to the discovery of ancient tunnels during contemporary mining works in various parts of Europe, such as Italy, Greece, and the Iberian Peninsula [19]. Among these researches, the work of the brothers Enrique (1857–1933) and Luis (1860–1934) Siret y Cels in southeast Spain should be remembered in particular. They, in addition to excavating prehistoric settlements, paid particular attention to metallurgical indications, analyzed some of the metal finds [20], and studied the ancient artifacts found inside the mines, such as the copper mine at Herrerías [21]. Despite the significant premises that developed between the nineteenth and the beginning of the twentieth century, the analytical results obtained on ancient metals reached the set goals only very partially. This was due to various factors. On the one hand, the available techniques consisted basically of wet analysis, a methodology that forced to sacrifice part, sometimes substantial, of the finds. This meant that the items examined were often made up of waste materials, without a recognizable shape, and therefore not typologically and chronologically recognizable. On the other hand, there was no real planning for the interventions to answer explicit historical questions. The essential cooperation between chemists and archaeologists occurred only occasionally. In addition to the aforementioned work by the Sirets, another example of this collaboration was carried out at the beginning of the century by the Italian scholar Angelo Mosso (1846–1910) [22], who examined the productive techniques of Italian and Cretan protohistoric bronzes (Fig. 24.4). He commissioned the chemical laboratory of the Regio Arsenale of Turin to carry out a systematic analysis of the archaeological objects that he previously selected. Mosso, however, was a completely anomalous character in the European landscape of those years, which combined both humanistic and scientific skills. He was an important physiologist (he invented the first neuroimaging technique), who shifted to archaeology in the last years of his life.

24.2.4 The New Approaches: Twentieth Century 24.2.4.1 Theoretical Archaeology and Archaeometallurgy Although archaeometallurgical research had a considerable development in the nineteenth century to have the first and most relevant use of the studies on ancient metals in theoretical archaeology, it was necessary to wait until the first half of the

680

C. Giardino

Fig. 24.4 Angelo Mosso (https://upload.wikimedia. org/wikipedia/commons/b/be/ AngeloMosso.jpg)

twentieth century, when the British scholar Vere Gordon Childe (1892–1957) put to good use the data on prehistoric metals elaborated up to that time [23–25]. His theories that social evolution and the formation of social classes were connected with development and technological specialization, particularly of metals, influenced most of the subsequent researchers [26–28]. According to Childe, metallurgy would have developed in the urban centers of the Near East due to the presence of an economic surplus and the desire of the elites to maintain their social status. Mechanisms of cultural interaction and dissemination (including the movement of itinerant metallurgists) would have allowed the transmission of technology from the Near East to “barbaric” areas, such as Europe. Childe, a full-fledged member of the historical-cultural school, was a declared follower of the materialist philosophy of Karl Marx (1818–1883). However, as regards technological studies, Childe was both materialistic and idealistic at the same time, as Marx also was. According to Childe, technologies are the expression of human thought [29], echoing the opinion of Marx and Engels that individuals coincide with their production, with what they produce and how they produce it [30]. Therefore, in the

24

Archaeometallurgy: A Discipline Between Past and Future

681

study of archaeological artifacts, we cannot limit ourselves to observing them, but we must also understand the processes that led to the production of objects, in their social context. To overlook this particular aspect led many Marxist scholars to create models of ancient socioeconomic dynamics that completely ignore the human aspects of technology and ancient commerce. Starting from the 1960s, New Archaeology (or Processual Archaeology) was the dominant cultural paradigm in archaeological studies, especially between the two sides of the Atlantic. The New Archaeology purported to procure data in support of its theories in an empirical and scientific manner, with the presumption of elaborating, likewise the “exact” sciences, universal laws specific to each culture through the exam of artifacts. Although archaeometry was increasingly regarded as a fundamental contribution to archaeological disciplines, the primary interest toward the elaboration of models to explain technological innovation [31: 296–297] often led to neglect the evidence that emerged from the studies on ancient technologies. This marginalized technological studies, and therefore archaeometallurgy, confined them to museum technicians and scientists. Theoretical developments were carried out by sectors such as the History of Science [32, 33]. The main theoretical contributions to archaeometallurgical studies drawn up in the second half of the twentieth century came from the Anglo-American School of Materials Science by Ronald Tylecote (1916–1990), Beno Rothenberg (1914–2012), Robert Maddin (1918–2019), and Cyril Stanley Smith (1903–1992) and from the French techno-anthropological school of “material processes” created by LeroiGourhan and his followers. André Leroi-Gourhan (1911–1986) [34] developed the idea of “technical-operational chain” (chaîne opératoire), which is the set of chained steps that occur in the production of artifacts. These start from the collection of raw materials to their abandonment, passing through the different manufacturing phases, their use, and their reconstruction, thus transforming the material into a cultural product. Significant applications of Leroi-Gourhan’s method to the field of metallurgy can be found in the work of Philippe Fluzin [35, 36]. According to the French school, the combination of technological choices made by people at a particular sociocultural phase in the production and use of material culture [37] can provide valuable data to understand the sociopolitical and the ideological structures of an ancient society [38, 39]. Tylecote’s and Stanley Smith’s studies [40–43] also had deep repercussions on archaeometallurgical studies which merged the scientific data coming from the Science of Materials with those provided by experimental archaeology and by archaeological field research. Stanley Smith’s statement “. . . Technology is more closely related to art than science - not only materially, but conceptually as well, because the technologist, like the artist, must work with many analogous complexities” [44: 325] paved the way for a profound interpretative revision of metal objects. These were the theoretical premises of the innovative Heather Lechtman’s studies on Andean metallurgy [45–48], in which she demonstrated how the metallurgical production of the pre-Columbian people – i.e., the realization of tumbaga, an alloy of copper, gold, and silver in varying proportions – was the direct reflection of their religious and social conceptions, hence of their

682

C. Giardino

Weltanschauung. The thought of the American sociologist and economic historian Immanuel M. Wallerstein (1930–2019) with his world-systems theory [49] had a significant influence. In the application of this model in archaeology, a dominant “core” region receives raw materials from a dominated “periphery” region, in exchange for finished products and/or subsistence goods in an interdependent relationship [50]. In archaeometallurgy, the world-systems theory was applied, among others, by Evgenij Nikolaevich Chernykh, based on typological, technological, and chemical analyses carried out by him on a considerable amount of metal finds. In his model, the “metallurgical province” – a geographically and chronologically defined region – is composed of “metallurgical focuses” (where the metal is excavated and smelted from its ores) and “metalworking focuses” (where the raw metal is imported, refined and finished); the “metallurgical focuses” are considered as “cores” while the “metalworking focuses” as dependent “peripheries” [51]. Some European institutions have devoted ample space to archaeometallurgy in its various aspects during the second half of the twentieth century, and numerous scholars have produced important syntheses by examining aspects such as metal production or iron technology [52–54]. The British Museum, thanks in particular to the studies of Paul T. Craddock, published relevant works in its Occasional Papers; also in the UK, the Historical Metallurgy Society, founded in 1962, produced many publications and publishes two relevant journals, Historical Metallurgy and The Crucible. In Germany, the Deutsches Bergbau-Museum in Bochum (founded in 1930, but which took on its current profile as a research center in the 1960s and 1970s) promotes international archaeometallurgical investigations – among which the studies conducted by Gerd Weisgerber (1938–2010) in the Middle East – and publishes authoritative studies in the journal Der Anschnitt and its supplement volumes (Beiheft).

24.2.4.2 A Systematic, Extensive Analytical Method In the twentieth century, and especially from the mid-1900s, ancient metals – particularly the prehistoric ones – became the object of relevant and extensive analytical investigations. The considerable increase in research was linked to the development of new analytical methodologies that replaced those of the nineteenth century conducted by wet assay. These technological innovations have allowed the realization of vast investigative projects that characterize this century, thanks to the spread of new instruments in the cultural heritage, such as the spectrometer, which has allowed to obtain accurate results in a short time and with small samples [55]. A first examples of this more modern scientific approach is represented by the studies conducted by C. H. Desch within the Sumerian Copper Committee; the numerous analyses conducted on both archaeological finds and minerals aimed at identifying the source of metals used in ancient Mesopotamia [56]. From the middle of the century, the research aimed at systematically investigating ancient alloys multiplied, especially by German scholars. After the Second World War, extensive research on European protohistoric metallurgy originated in Germany, carried out by H. Otto and W. Witter [57]. Some years later, with the support of three important German institutions, the Württembergischen Landesmuseum of Stuttgart, the Deutsche Forshungsgemeinshaft, and the Römish-Germanischen Zentralmusuems

24

Archaeometallurgy: A Discipline Between Past and Future

683

of Mainz, an ambitious program of compositional spectroscopic analysis was developed, using about 22,000 pre-protohistoric metal artifacts coming from all over Europe. The aim was to determine the origin of copper based on trace elements and to study the evolutionary development of European prehistoric metallurgy [58–60]. Unfortunately, none of the two objectives were achieved; nevertheless, this remains one of the largest corpora of analysis and it can be still cited with confidence. Despite the failure to achieve the expected objectives, the vast program carried out in Germany acted as a stimulus for similar projects in other countries. In this regard, the research carried out on the ancient metals of Scandinavia [61] and those, of very vast proportions, carried out on prehistoric metal finds from the regions of the vast territory of the Soviet Union should be mentioned. With this last project, about 60,000 bronzes were analyzed, to study the development of ancient mining and metallurgy in the northern half of Eurasia [62]. With the recognition that chemical analysis cannot determine the geological origin of the metal used to make archaeological metal objects, systematic researches were carried out with a different analytical technique, the measurement of lead isotope ratios in the copper. A basic contribution to the studies on the prevalence of copper used in the production of ancient objects was provided by the decades-long work carried out at the Isotrace Laboratory of the University of Oxford between 1975 and 2002 by Noël H. Gale (1931–2014) and by Zofia Stos-Gale, creating the Oxford lead isotope database (OXALID) [63, 64]. A good example of a concrete answer to the question about the origins of copper ores is the solution to the old query about the origin of the ox-hide ingots of the Bronze Age [65, 66].

24.2.5 The Recent Developments in Archaeometallurgical Research: An Outlook for the Future (End of the Twentieth Century to Beginnings of the Twenty-First Century) Between the end of the 1990s and the beginning of the new century, the final results of the large Spanish national project were published; the “Proyecto Arqueometallurgia,” was coordinated by Josè María Cabrera and Salvador Rovira. The project, which began in the 1980s, analyzed the first metallic evidence of the Iberian Peninsula from an archaeological and analytical perspective [67, 68]. The study made use of an integrated system of investigative techniques (including SEM, EDS, classical metallography, X-ray fluorescence) on a vast number of artifacts. It provided relevant insights that affected not only the history of Iberian metallurgy but also the European and extra-European one. The current trends of archaeology – and consequently of archaeometallurgy – tend to investigate aspects related to the study of identities (such as age, ethnicity, gender, class, nationality, social boundaries, etc.) and their effects on the material culture [69–73]. Starting from the last decades of the twentieth century, studies on non-European metallurgies have been developed, such as those of sub-Saharan Africa, in the past the object of ethnographic investigations [74], and those of the Arabian peninsula [75–78], the Levant [79], and Iran [80–82]. The vast territory of China is another

684

C. Giardino

area often neglected in the past, although of great importance. Here the research has investigated, among other things, ancient mining and smelting sites, as well as the first Chinese metallurgy, its native, or imported origin [83–88]. In prehistoric studies – traditionally the most focused on archaeometallurgical investigations – interest has often concentrated, for the twentieth century, on the chronology of metal finds, leaving out other important aspects, such as the determination of operating chains [89]. Only in recent times, the radiocarbon datings of archaeological complexes have been applied systematically in many regions; the chronology of finds has been mostly determined on a stylistic-typological basis. In the study of prehistoric metal finds, the most significant expression of the chrono-typological approach was the “Prähistorische Bronzefunde” (PBF) project, started in 1966, which led to the publication of over 140,000 European and Asian items in more than 180 publications. Organized on a regional basis and by classes of artifacts, sometimes also including analytical data. This approach has allowed reviewing technological and social processes of model reception and to re-examine the transmission routes of the metallurgical technology. Studies of ancient mines and slag have multiplied in recent years [90–93] (Figs. 24.5 and 24.6). However, investigations on slag are still scarce, especially those related to lead, silver, and gold. Moreover, in archaeological publications, the

Fig. 24.5 Al-Moyassar (Sultanate of Oman). The collapsed entrance of a Bronze Age mine (photograph by C. Giardino)

24

Archaeometallurgy: A Discipline Between Past and Future

685

Fig. 24.6 Bilad Al-Maidin (Sultanate of Oman). Prehistoric tapped slag on the surface of a slag heap (photograph by C. Giardino)

discoveries of slag connected with different metallurgical processes (i.e., smelting, casting, and smithing) are often misunderstood and confused. This gap may be corrected in the future with more widespread dissemination of archaeometallurgical studies on slag (Fig. 24.7). An example of these new trends is in the Alpine area and in Italy, where these investigations have already made it possible to ascertain the great antiquity of regulating the copper ore smelting processes, including the furnace temperature and the internal atmosphere [94, 95]. In the early years of the new century, some summary books appeared, like manuals, aimed to assess the state of studies, also for university teaching purposes [96–99]. At this time the amount of metallurgical data from different world regions had dramatically increased, making a synthesis of the early development of metallurgy a difficult and complex matter [100]. In research, the contribution of experimental archaeology and ethnoarchaeology is today increasing systematically; they provide important insights into knowledge and validations (or denials) to the hypotheses formulated based on archaeological and analytical data, particularly as regards the reconstruction of ancient production techniques (Figs. 24.8 and 24.9).

24.3

Analytical Techniques and Methodologies

24.3.1 General Requirement for Archaeometallurgical Investigations The work of archaeologist (and in many respects also of art historian) is to reconstruct the historical and sociocultural phenomena starting from the few material remains connected with the material activities of man. Hence there is the need to resort with increasing frequency to the contribution of analytical data that allows clarifying otherwise obscure or uncertain aspects. From a methodological point of

686

C. Giardino

Fig. 24.7 Malegno (Brescia, Italy). Prehistoric slag (photograph by C. Giardino)

Fig. 24.8 Experimental archaeology. Casting a Bronze Age sword (photograph by C. Giardino)

view, it is necessary to premise that any test and analysis carried out on archaeological and historical-artistic artifact must necessarily provide an answer to specific problems, previously explained by the researchers. This requirement is present both if the questions concern a better knowledge and a more precise cultural and technological framing and if they concern conservation and restoration. The methodology to be applied must therefore be chosen from time to time based on the questions to be

24

Archaeometallurgy: A Discipline Between Past and Future

687

Fig. 24.9 Ethnoarchaeology. Lost-wax casting at Mojoagung (East Java, Indonesia): the hollow wax copy of the original model is removed from the mold (photograph by C. Giardino)

answered: the analytical techniques will be selected from those most suitable to achieve the target. Because archaeometallurgy is a typically multidisciplinary research field, it requires the active contribution of different expertise and knowledge. The investigations usually require close collaboration between specialists from different disciplines, both scientific and archaeological-artistic, since specific skills are required to carry out the research and to interpret the results. In archaeometallurgical studies, it is necessary to confirm the results obtained with a technique and to compare them with the results obtained from a different technique, based on different physical principles. This procedure allows gaining a mutual validation, or at least a greater understanding of the obtained information. The choice of methodology to be adopted in the research must be evaluated according to the technical characteristics of the object to be studied, in the environment that hosts it, and in accordance with the final aims of the research. A good program of archaeometallurgical analysis must answer some fundamental questions, such as the provenance of the raw materials, the smelting techniques, the metal composition, its possible alloying and the methods used to obtain it, the manufacturing processes of finished objects, and the possible identification of workshops. Today we have very precise and sensitive investigative instruments, capable of detecting elements present in a metal product in quantities lower than a part per million (ppm), as well as measuring parameters such as hardness, resistance, and flexibility. However, it is necessary to keep in mind that not all analytical techniques used on an industrial level can be applied to archaeometallurgy, as this must take into account the uniqueness that characterizes cultural heritage. Only a

688

C. Giardino

limited number of objects can sometimes be “sacrificed” with destructive investigations, such as minerals or slags [101]. However, consider that today, unlike in the past, even most invasive investigations require only a few milligrams of a sample, which can be obtained without causing real damage by drilling a hole of tenths of millimeters.

24.3.2 Nondestructive and Noninvasive Investigations The study of archaeological and artistic metals is a preferred application in the research field for nondestructive and noninvasive investigations, because of the uniqueness and unrepeatability that characterize the examination of such artifacts. These techniques, widely applied in industry, are a system for monitoring and studying the structural model of archaeological items and works of art. They allow the examination of one or more physical-chemical characteristics, without altering the object being investigated, as would be the case if samples were taken. Therefore, this type of examination meets the basic requirements of protection and conservation set for these particular materials. At the end of the test, the objects have no trace of the intervention: an accurate record of all the analytical parameters is therefore indispensable. The number and precise location of the analyzed points must therefore be documented, as well as, obviously, the description of the techniques, methodologies, and the adopted equipment, to ensure that future investigations can be replicated. Noninvasive exams, while producing minor and localized modifications in very limited areas of the object, do not affect the conservation of the piece, nor alter its enjoyment. An example is the preventive removal of a small area of corrosion that often covers archaeological finds in copper alloys to carry out X-ray fluorescence analyses (XRF). Usually, nondestructive tests are different from nondestructive analyses. Nondestructive tests provide assessments and opinions on the structural modifications of an artifact, such as its mechanical behavior in the presence of external stresses. On the other hand, nondestructive analyses give quantitative information on the materials of which the object is made: they can provide its elemental composition or information about the patinas (natural or artificial) that cover it. The same techniques may, from time to time, assume the character of nondestructive or noninvasive analyses, depending on the examined objects, as with X-ray fluorescence applied to the study of the elemental composition of metals. Even the study with a scanning electron microscope (SEM) can be part of nondestructive tests, provided that the dimensions of the object are compatible with those of the instrument chamber where it is to be examined.

24.3.2.1 Noninvasive Compositional Analyses Energy-dispersive X-ray fluorescence (ED-XRF), an emission spectroscopy technique, is one of the most powerful tools available today to carry out qualitative and quantitative analyses on artistic and archaeological metals in a nondestructive or

24

Archaeometallurgy: A Discipline Between Past and Future

689

noninvasive manner. Recent improvements in the instrumentation allow obtaining very high sensitivity and accuracy. The analyses can be carried out directly where the materials to be examined are stored, thanks to the widespread use of portable instruments, considerably facilitating its use. XRF allows to carry out many measurements on different areas of the same piece, without extracting any sampling, thus highlighting the compositional dissimilarities that normally characterize ancient artifacts [see, among others, 102–104]. The reduced penetration capacity of the radiation used makes it possible to examine only the surface of the objects. It is, therefore, necessary to pre-clean the spot where the measurement is to be carried out, thus removing the surface corrosion that would otherwise affect the elemental composition of the metal (Figs. 24.10 and 24.11). On the other hand, the precise knowledge of the composition of the superficial layers of the object is very useful for conservation and restoration, to prepare specific and targeted methods of protection against degradation. In the noble metals, such as gold, the absence of any superficial alteration levels allows a nondestructive use of XRF. XRF is also useful to investigate evidence of metallurgy as crucibles and molds [105, 106] (Figs. 24.12 and 24.13).

24.3.2.2 Radiological Techniques The radiographic techniques have particular relevance in nondestructive techniques. They are based on the capability of X-rays (Röntgen) to pass through many materials, thanks to their reduced wavelength. They give a visible image of

Fig. 24.10 Small Bronze Age tool. A small spot has been cleaned for the XRF measurement (red arrowhead), removing the surface patina (photograph by C. Giardino)

690

C. Giardino

Fig. 24.11 XRF spectrum of a prehistoric bronze (courtesy G. Paternoster)

the internal structure based on the different types of materials, radio-opaque to a different extent. X-rays show fractures, ancient restorations, casting defects, and weldings. They also allow to obtain data on the working techniques used during manufacturing. This examination can be indispensable before the restoration of highly corroded objects since it detects eventually hidden decorations: this is the case of silver inlay or niello artifacts, or iron objects, covered with thick layers of patina, whose removal, without preventive X-rays, can lead to the removal of decoration [107]. The development of traditional X-ray imaging is 3D computed tomography: tomographic data can be displayed as cross sections or as 3D images, allowing the inner inspection of the object [108]. Similar to X-rays, investigations are those carried out with γ rays that produce images similar to the radiographic ones. The γ rays have a greater capability to penetrate the materials in comparison with X-rays; this makes the γ rays very useful in investigating particularly thick metal objects, such as bronze statues.

24.3.2.3 Nondestructive Testing: Ultrasonics, Acoustic Emission, and Thermography Ultrasonics are usefully employed in diagnostics; ultrasonics are sound waves with frequencies above 20 kHz and therefore above the threshold values of the human ear.

24

Archaeometallurgy: A Discipline Between Past and Future

691

Fig. 24.12 Fragment of clay mold for metalware production (Castro, Lecce, fourth to third century BC) (photograph by C. Giardino)

Beams of ultrasonic waves are generated to record how they are modified (reflected, refracted, or absorbed) by their encounter with an object; the examination of how the object interferes with the ultrasounds gives information on the structural state of the object itself. This technique finds an ideal application in the investigations of metals [109: 238–251]. Ultrasonics allows analyzing not only thickness variations but also internal dissimilarities such as cracks and bubbles due to casting defects. It also allows identifying artifact elements made of different metals, such as pins, welds, macro inclusions, and plugs. Another relevant nondestructive test is the acoustic emission, which detects and measures the release of elastic energy of a structure when it is subjected to slight stresses from the environment. The daily thermal cycles produce microstructural

692

C. Giardino

Fig. 24.13 XRF analysis of a Bronze Age crucible (photograph by C. Giardino)

modifications in materials, with the emission of signals, which can diagnose dissimilarities, defects, and micro-cracks. This technique is used mainly (but not only) in studies of environmental degradation of outdoor monumental complexes, where it highlights mechanical and static problems and analyzes chemical and electrochemical corrosion. An example of the application are the studies carried out on the Roman bronze statue of Marcus Aurelius in Rome [110]. Thermography is based on the physical principles of heat transmission by radiation. It allows viewing the temperature distribution on the surface of an object highlighting any structural discontinuities on the surface and, therefore, differentiating inserts and repairs. A significant example of the use of this technique on ancient finds is the study conducted by high-resolution thermography on the bronze Etruscan sculpture of the Chimera of Arezzo [111].

24.3.3 Minimally Invasive Investigations Techniques such as spectrometry and metallography have been used for a long time. Unlike in the past, modern analytical techniques allow taking samples of very small dimensions. Therefore we can speak of minimally invasive analyses for a whole series of investigations that can be performed on ancient metals.

24.3.3.1 Metallography Metallographic examination generally requires a small sample of the specimen to be placed under the metallographic microscope. The study is carried out after careful preparation, with polishing and acid etching. Metallography on metal items is based on the microscopic study of the deformations undergone during the working

24

Archaeometallurgy: A Discipline Between Past and Future

693

processes by the crystals of which the metals are made. The appearance of the crystalline structure, therefore, varies according to whether an object was only cast, or was also hammered, or annealed [112] (Figs. 24.14 and 24.15). In addition to the metallographic structure, it is possible to observe inclusions, casting defects, cavities, and corrosion phenomena. The optical metallography is juxtaposed by the electronic one – which is generally interfaced with the EDS microprobe – which allows the examination to a much greater resolution and, thanks to the EDS, to simultaneously have the microanalysis of the observed structures (Figs. 24.16 and 24.17). Although nondestructive metallography is being used, it is still not very widespread. A point on the find is subjected to polishing and etching: then the intact piece is placed under the microscope, without taking samples. The latter technique goes well with XRF, as the already cleaned point for the measurement is the one subjected to metallography. Metallography also allows the investigation, under normal and polarized light, of slag, a complex mixture of many constituents, basically, silicates, that are the waste results of the metallurgical operations [113, 114].

24.3.3.2 Microinvasive Compositional Analyses The spectrometric techniques require a micro-sampling from the piece; of particular importance is mass spectrometry (MS), which allows obtaining quantitative analyses with very high sensibility and precision.

Fig. 24.14 Polished and etched section of a Roman bronze, showing a dendritic structure: the piece was only cast (50) (photograph by C. Giardino)

694

C. Giardino

Fig. 24.15 Polished and etched section of a Roman bronze, showing twinned grains: the piece was annealed after cold working (200) (photograph by C. Giardino)

MS has relevance in the determination of lead isotopic ratios. Their composition is the main source of information on the provenance of the ores used in ancient metal production. Lead isotopes are analyzed using a multi-collector-inductively coupled plasmamass spectrometer (ICP-MS). Lead is normally associated with copper, silver, iron, and zinc, although in small quantities. It is composed of four isotopes: 204Pb, 206Pb, 207Pb, and 208Pb. Each mineral ore has a peculiar Pb isotopic composition, which depends on the ore deposit age and the presence of lead and, possibly, uranium and thorium. Therefore the proportions of the four isotopes in each ore deposit are linked to the geochemical history of the deposit itself. Unlike the chemical composition, the ratio between the lead isotopes is relatively constant within the same ore deposit, and, above all, it is not influenced by metallurgical processes: therefore the original mineral, the resulting metal, and the smelting slag have the same isotopic imprint [115–122]. By comparing the isotopic compositions of lead minerals from known deposits and those observed in manufactured products, it is, therefore, possible to identify the origin of the metal. The most common system to present isotopic relationships is to report the lead isotopic ratios as two images in two-dimensional diagrams [123]. Information on the isotopic composition of lead in those deposits are related to archaeological and historical-artistic research that can be obtained from databases published in the literature and on dedicated Internet sites.

24

Archaeometallurgy: A Discipline Between Past and Future

695

Fig. 24.16 SEM image of a Roman bronze (photograph by C. Giardino)

Remarkable use in archaeological studies has the laser ablation and inductivelycoupled plasma mass spectrometry (LA-ICP-MS). The laser ablation sampling technique uses the irradiation of a laser beam, and then the laser-induced sample aerosols and vapors are introduced to the ICP. Recently, portable devices have been developed to carry out on-site investigation [124]. The neutron activation analysis (NAA) is an analytical technique based on nuclear processes that allow obtaining qualitative and quantitative information on the elements which are present, by measuring their radioactivity. It is a chemicalanalytical method of high sensitivity that provides data on trace elements present in metal alloys. Atomic absorption spectroscopy (AAS) and atomic emission spectroscopy (AES) are the spectroanalytical techniques for detecting the quantitative determination of chemical elements in a liquid solution through the application of characteristic wavelengths of electromagnetic radiation from a light source. They have very high sensitivity, between p.p.m. (parts per million) and p.p.b. (parts per billion). X-ray diffractometry (XRD) determines the crystalline phases contained in the minerals. In the case of metals, it finds particular use in identifying changes in the surfaces of ancient items due to degradation, such as patinas.

696

C. Giardino

Fig. 24.17 SEM-EDS elemental micro-mapping of a prehistoric bronze (photograph by C. Giardino)

References 1. Thompsen CJ (1836) Ledetraad til Nordisk Oldkyndighed. S.L. Møllers bogtr, Kjöbenhavn 2. Dizé M, D’Arcet M (1790) Analyse du cuivre, avec lequel les Anciens fabriquoient leurs Médailles, les Instrumens tranchant, Observations et Mémoirs sur la physique, sur l'histoire naturelle et sur les arts et métiers. Journal de physique, de chimie, d'histoire naturelle et des arts XXXVI, I, 272–276, Avril, Paris 3. Wiegleb JC (1778) Chemisch Untersuchung einiger künstlichen Metallarten, woraus verschiedene aus dem Altertum herrührende Instrumente versertiget gewesen, welche im vorigen Jahre in einer benachbarten Gegend gefunden worden find, Acta Academiae Electoralis Moguntinae Scientarium Utilium, quae Erfurti est ab annum MDCCLLVII, 50–57, Erfurti 4. Hjelm PJ (1797) Om konsten at härda Koppar, Kongl. Vetenskaps Academiens nya handlingar för Mon. Apr.–Juni XVIII, 98–110 5. Mongez A (1800) Mémoires sur le bronze des Anciens, Mémoires de l'Institut national des sciences et arts. Littérature et beaux-arts 5:187–516 6. Pownall G (1786) An account of some Irish antiquities, Archaeologia, or Miscellaneous tracts relating to antiquity, III, 355–356 7. Klaproth MH (1798) Mémoire de numismatique docimastique, Mémoires de l’académie Royale des Sciences et Belles-lettres depuis l’avénement de Fréderic Guillaume II au Trône, Classe de Philosophie Expérimentale, 97–113, Berlin

24

Archaeometallurgy: A Discipline Between Past and Future

697

8. Becquerel M (1832) De la Cristallisation de quelques Oxides métalliques. Annales de Chimie et de Physique 51:101–106 9. Hünefeld L, Picht F (1827) Rügens metallische Denkmäler der Vorzeit, vorzugsweise chemisch bearbeitet, und als Beytrag zur vaterländischen Altertumskunde. Leopold Voss, Leipzig 10. Hausmann JFL (1815) Über die Eisendarstellungsprozesse der Alten. In: Schweigger JSC (ed) Journal für die Chemie und Physik 14. Schrag’sche Buchhandlung, Nürnberg, pp 247–261 11. Roloff EH (1810) Ueber den Bergwesen und die Metallurgie des alten Spaniens. In: Gehlen AF (ed) Journal für die Chemie, Physik und Mineralogie IX. Realschulbuchhandlung, Berlin, pp 609–652 12. Boeckh A (1818) Über die antiken laurischen Silberbergwerke in Attika, Abhandlungen der Preußischen Akadademie der Wissenshaften: Historisch-Philologische Klasse, 85–140 13. Cordellas A (1869) Le Laurium. de Cayer, Marseilles 14. Mallett JW (1852) Account on a chemical examination of the celtic antiquities in the collection of the Royal Irish Academy Dublin. University Press, Dublin 15. Mallett JW (1855) Report on the chemical examination of antiquities from the Museum of the Royal Irish Academy. Trans R Irish Acad 22:313–342, V, Science 16. Göbel F (1842) Über den Einfluß der Chemie auf die Ermittlung der Völker der Vorzeit oder Resultate der chemischen Untersuchung metallischer Alterthümer insbesondere der in den Ostseegouvernements vorkommenden, behufs der Ermittlung der Völker, von welchen sie abstammen. Ferdinand Enke, Erlangen 17. Dixon C (1831) Some accounts of the lead mines of Ajmer. Gleanings Sci 28:111 18. Percy J (1870) The metallurgy of lead. John Murray, London 19. Simonin ML (1858) De l’exploitation des mines et de la métallurgie en Toscane pendant l’Antiquité et le Moyen Age. Annales des Mines V XIV:557–615 20. Siret E, Siret L (1890) Las Primeras Edades del Metal en el Sudeste de España. Resultados obtenidos en las excavaciones hecas por los autores. Barcelona 21. Siret L (1909) Villaricos y Herrerías. Antigüedades púnicas, romanas, visigóticas y árabes. Memoria descriptiva e histórica (Establecimiento Tipográfico de Jaime Ratés, Madrid, 1907). Idem: Memorias de la Real Academia de la Historia, XIV, 380–478, Madrid 22. Mosso A (1908) Le armi più antiche di rame e di bronzo, Memorie della R. Accademia dei Lincei, S. 5A, Classe di scienze morali, storiche e filologiche, XII. Accademia dei Lincei, Roma 23. Childe VG (1930) The Bronze age. Cambridge University Press, Cambridge 24. Childe VG (1944) Archaeological ages as technological stages. J R Anthropol Inst G B Irel 74: 7–24 25. Childe VG (1958) The prehistory of European society. Harmondsworth, London 26. Rowlands MJ (1971) The archaeological interpretation of prehistoric metalworking. World Archaeol 3:210–224 27. Trigger BG (1986) The role of technology in V. Gordon Childe’s archaeology. Nor Archaeol Rev 19:1–14 28. Wailes B (1996) V. Gordon Childe and the relations of production. In: Wailes B (ed) Craft specialization and social evolution: in memory of V. Gordon Childe, University Museum monograph 93, University Museum of Archaeology and Anthropology. University of Pennsylvania, Philadelphia, pp 3–14 29. Childe VG (1956) Society and knowledge. Harper and Brothers, New York 30. Marx K, Engels F (1932) Die Deutsche Ideologie. Manuscript: First edition Marx-EngelsLenin-Institut, Moskau, 1845–46 31. Trigger BG (1989) A history of archaeological thought. Cambridge University Press, Cambridge 32. Pfaffenberger B (1992) Social anthropology of technology. Annu Rev Anthropol 21:491–516 33. Pfaffenberger B (1999) Worlds in the making: technological activities and the construction of intersubjective meaning. In: Dobres MA, Hoffman CR (eds) The social dynamics of

698

C. Giardino

technology: practice, politics, and world views. Smithsonian Institution Press, Washington, DC, pp 147–166 34. Leroi-Gourhan A (1964) La Geste et la Parole I: Technique et Langage. Éditions Albin Michel, Paris 35. Fluzin P (1999) Il processo siderurgico: evoluzione storica e indizi archeometrici. In: Cucini Tizzoni C, Tizzoni M (eds) La miniera perduta. Cinque anni di ricerche archeometallurgiche nel territorio di Bienno. Tipografia Camuna, Breno, pp 61–92 36. Fluzin P, Benoit P, Kienon HT, Kiethega JB, El Kedim O (1995) Apports de l’archéométie à la restitution de la chaîne opératoire des procédés sidérurgiques directs à partir des vestiges archéologiques; intérêtsdes comparaisons ethnoarchéologiques. In: Magnusson G (ed) The importance of ironmaking. Technical innovation and social change, Papers presented at the Norberg Conference on May 8–13, 1995. Jernkontorets Bergshistoriska Utskott, Stockholm, pp 56–64 37. Lemonnier P (1986) The study of material culture today: towards an anthropology of technical systems. J Anthropol Archaeol 5:147–186 38. Dobres MA, Hoffman CR (1994) Social agency and the dynamics of prehistoric technology. J Archaeol Method Theory 1(3):211–258 39. Stark MT (1998) Technological choices and social boundaries in material culture patterning: an introduction. In: Stark MT (ed) The archaeology of social boundaries. Smithsonian Institution Press, Washington, DC, pp 1–11 40. Tylecote RF (1962) Metallurgy in archaeology. Edward Arnold, London 41. Tylecote RF (1987) The early history of metallurgy in Europe. Longman, London/New York 42. Smith CS (1970) Art, technology, and science: notes on their historical interaction. Technol Cult 11:493–549 43. Smith CS (1977) Metallurgy as a human experience: an essay on man’s relationship to his materials in science and practice throughout history. American Society for Metals, Metals Park 44. Smith CS (1981) On art, invention, and technology. In: Smith CS (ed) A search for structure. MIT Press, Cambridge, MA, pp 325–331 45. Lechtman H (1979) Issues in Andean metallurgy. In: Benson EP (ed) Pre-Columbian metallurgy of South America. Dumbarton Oaks, Washington, DC, pp 1–40 46. Lechtman H (1984) Andean value systems and the development of prehistoric metallurgy. Technol Cult 25(1):1–36 47. Lechtman H (1993) Technologies of power: the Andean case. In: Henderson J, Netherly P (eds) Configurations of power in complex societies. Cornell University Press, Ithaca, pp 244–280 48. Lechtman H (1999) Afterword. In: Dobres M-A, Hoffman CR (eds) The social dynamics of technology. Smithsonian Institution Press, Washington, DC, pp 223–232 49. Wallerstein I (1974) The modern world-system: capitalist agriculture and the origins of the European world-economy in the sixteenth century, 1. Academic Press, New York 50. Algaze G (1993) The Uruk world system. University of Chicago Press, Chicago 51. Chernykh EN, Avilova LI, Orlovskaja LB (2000) Metallurgical provinces and radiocarbon chronology. Rossijskaja Akademija Nauk, Institut Arkheologii, Moscow 52. Craddock PT (1995) Early metal mining and production. Edinburgh University Press, Edinburgh 53. Pleiner R (2006) Iron in archaeology. Early European Blacksmiths. Archeologický ústav AVČR, Praha 54. Mohen J-P (1990) Métallurgie préhistoriques. Introduction à la paléométallurgie. Masson, Paris/Barcelone/Mexico 55. Otto H (1940) Die Anwendung der Spektranalyse für kulturhistorishen Fragen. Spectrochim Acta 1:381–399 56. Detsch CH (1928–1938) Reports of the Sumerian Committee, Reports of the meetings of the British Association for the Advancement of Science. Office of the British Association, London 57. Otto H, Witter W (1952) Handbuch der ältesten vorgeschichtlichen Metallurgie in Mitteleuropa. J. Ambrosius Barth Verlag, Leipzig

24

Archaeometallurgy: A Discipline Between Past and Future

699

58. Junghans S, Sangmeister E, Schröder M (1960) Studien zu den Anfängen der Metallurgie. Metallanalysen kupferzeitlicher und frübronzezeitlicher Bodefunde aus Europa 1. Gebr. Mann Verlag, Berlin 59. Junghans S, Sangmeister E, Schröder M (1968) Studien zu den Anfängen der Metallurgie. Kupfer und Bronze in der frühen Metallzeit Europas 2, 1–3. Gebr. Mann Verlag, Berlin 60. Junghans S, Sangmeister E, Schröder M (1974) Studien zu den Anfängen der Metallurgie. Kupfer und Bronze in der frühen Metallzeit Europas 2, 4. Gebr. Mann Verlag, Berlin 61. Culberg C (1968) On artifact analysis. Acta Archaeologica Lundensia 4, 7. Gleerup, Lund 62. Chernykh EN (1992) Ancient metallurgy in the USSR. The early metal age. Cambridge University Press, Cambridge 63. Stos-Gale ZA, Gale NH (2009) Metal provenancing using isotopes and the Oxford archaeological lead isotope database (OXALID). Archaeol Anthropol Sci 1:195–213 64. https://oxalid.arch.ox.ac.uk/. Accessed 16 Oct 2019 65. Gale NH, Stos-Gale ZA (2000) Lead isotope analysis applied to provenance studies. In: Ciliberto E, Spoto G (eds) Modern analytical methods in art and archaeology. Wiley, New York, pp 503–584 66. Hauptmann A (2009) Lead isotope analysis and the origin of Sardinian metal objects. In: Lo Schiavo F, Muhly JD, Maddin R, Giumlia-Mair A (eds) Oxhide ingots of the central Mediterranean. CNR – Istituto per gli Studi Micenei ed Egeo-Anatolici, Roma, pp 499–514 67. Rovira S, Montero I, Consuegra S (eds) (1997) Las primeras etapas metalúrgicas en la península Ibérica. I. Análisis de materiales. Instituto Universitario Ortega y Gasset, Madrid 68. Rovira S, Gòmez Ramos P (eds) (2003) Las primeras etapas metalúrgicas en la península Ibérica. III. Estudio metálograficos. S.n., Madrid 69. Shennan S (ed) (1989) Archaeological approaches to cultural identity. Unwin Hyman, London 70. Wright RP (ed) (1996) Gender and archaeology. University of Pennsylvania Press, Philadelphia 71. Sørensen M-LS (2000) Gender archaeology. Polity Press, Cambridge 72. Meskell L (2001) Archaeologies of identity. In: Hodder I (ed) Archaeological theory today. Polity Press, Cambridge, pp 187–213 73. Thornton CP, Giardino C (2008) Alla ricerca di un paradigma archeometallurgico. Rivista di Scienze Preistoriche LVIII:385–404 74. Bocoum H (ed) (2004) The origins of iron metallurgy in Africa. New light on its antiquity: West and Central Africa. UNESCO Publishing, Paris 75. Weeks LR (2003) Early metallurgy of the Persian Gulf. Technology, trade, and the Bronze Age world. Brill, Boston 76. Weisgerber G (2008) Metallurgy in Arabia. In: Selin H (ed) Encyclopaedia of the history of science, technology, and medicine in non-western cultures. Springer, Dordrecht, pp 1613–1622 77. Yule PA, Gernez G (eds) (2018) Early iron age metal-working workshop in the empty quarter, al-Ẓāḥira Province, Sultanate of Oman, Universitätsforschungen zur Prähistorischen Archäologie 316. Halbert-Verlag, Bonn 78. Giardino C (2019) Magan – the land of copper. Prehistoric metallurgy of Oman. Archaeopress Publishing Ltd, Oxford 79. Hauptmann A (2000) Zur frühen Metallurgie des Kupfers in Fenan/Jordanien. Der Anschnitt Beiheft 11. Deutsches Bergbau-Museum, Bochum 80. Hauptmann A, Rehren T, Schmitt-Strecker S (2003) Early bronze age copper metallurgy at Shar-I-Sokhta (Iran) reconsidered. In: Stöllner T, Körlin G, Steffens G, Cierny J (eds) Man and mining – Mensch und Bergbau: studies in honour of Gerd Weisgerber on occasion of his 65th birthday. Der Anschnitt, Beiheft 16. Deutsches Bergbau-Museum, Bochum, pp 197–213 81. Artioli D, Giardino C, Guida G, Lazzari A, Vidale M (2005) On the exploitation of copper ores at Shahr-i Sokhta (Sistan, Iran) in the 3rd millennium BC. In: Franke-Vogt U, Weisshaar H-J (eds) South Asian archaeology 2003. Proceedings of the seventeenth international conference of South Asian Archaeologists (Bonn 2003). Linden Soft Verlag, Aachen, pp 179–184

700

C. Giardino

82. Keykhaei M, Valiei MH, Shirazi R, Nadooshan FK (2012) Shahr-i Sokhta and the Bronze production workshop: a review, Interdisciplinaria archaeologica. Nat Sci Archaeol III(2):203–210 83. Bernard N (1980) The origins and development of metallurgy in China, with special reference to the crucible. In: Loofs-Wissowa HHE (ed) The diffusion of material culture, Asian and Pacific archaeology series 9. University of Hawaii, Honolulu, pp 215–263 84. Zhimin A (1981) Zhongguo zaoqi tongqi de jige wenti [Some problems concerning China’s early copper and bronze artifacts]. Kaogu Xuebao [Acta Archaeol Sin] 3:269–285 85. Shuyun S, Rubin H (1981) Gansu zaoqi tongqi de chubu yanjiu [A preliminary study of early Chinese copper and bronze artifacts]. Kaogu Xuebao 2 [Acta Archaeol Sin], 287–301. Translated in English in: Early China 9–10, 261–289 (1983–85) 86. Yangxiang L (1993) Zhongtiaoshan gutong kuangye yizhi chubu kaocha yanjiu [A preliminary study of ancient copper mining site at the Zhongtiaoshan region]. Q J Cult Relics 2 87. Jinjun M, Yanxiang L (2003) Early copper technology in Xinjiang, China: the evidence so far. In: Craddock P, Lang J (eds) Mining and metal production through the ages. British Museum Press, London, pp 111–121 88. Mei J, Wang P, Chen K, Wang L, Wang YC, Liu Y (2015) Archaeometallurgical studies in China: some recent developments and challenging issues. J Archaeol Sci 56:221–232 89. Ottaway BS (2001) Innovation, production and specialization in early prehistoric copper metallurgy. Eur J Archaeol 4(1):87–112 90. Rieser B, Schrattenthaler H (2004) Prähistorischer Kupferbergbau im Raum Schwaz/Brixlegg (Nordtirol). Geländebefunde und experimentelle Untersuchungen zur Schlägelschäftung. In: Weisgerber G, Goldenberg G (eds) Alpenkupfer – Rame delle Alpi, Der Anschnitt Beiheft 17. Deutsches Bergbau-Museum, Bochum, pp 75–94 91. Höppner B, Barteheim M, Huijsmans M, Krauss R, Martinek K-P, Pernicka E, Schwab R (2005) Prehistoric copper production in the Inn Valley (Austria), and the earliest copper in Central Europe. Archaeometry 47(2):293–315 92. Huijsmans M, Krauß R, Stibich R (2004) Prähistorischer Fahlerzbergbau in der Grauwackenzone. Neolithische und bronzezeitliche Besiedlungsgeschichte und Kupfermetallurgie im Raum Brixlegg (Nordtirol). In: Weisgerber G, Goldenberg G (eds) Alpenkupfer – Rame delle Alpi, Der Anschnitt Beiheft 17. Deutsches Bergbau-Museum, Bochum, pp 53–62 93. Bourgarit D, Rostan P, Burger E, Carozza L, Benoît M, Artioli G (2008) The beginning of copper mass production in the Alps: the Saint-Véran mining area reconsidered, historical metallurgy. J Hist Metallurgy Soc 42(1):1–11 94. Artioli G, Angelini I, Burger D, Bougarit E, Colpani F (2007) Petrographic and chemical investigations of the earliest copper smelting slags in Italy: towards a reconstruction of the beginning of copper metallurgy. In: Archaeometallurgy in Europe. Proceedings of the second international conference, Aquileia 2007. AIM, Milano, pp 12–20 95. Bourgarit D (2007) Chalcolithic copper smelting. In: La Niece S, Hook D, Craddock P (eds) Metals and mines: studies in archaeometallurgy. Duncan R. Hook, London, pp 3–14 96. Hauptmann A (2007) The archaeometallurgy of copper: evidence from Faynan, Jordan. Springer, NewYork/Heidelberg/Dordrecht/London 97. Giardino C (2010) I metalli nel mondo antico. Introduzione all’archeometallurgia. Nuova edizione aggiornata e ampliata. Laterza, Roma-Bari 98. Montero Ruiz I (2010) Manual de arqueometalurgia. Ilustre Colegio de Doctores y Licenciados en Filosofía y Letras y en Ciencias de la Comunidad de Madrid, Madrid 99. Roberts BW, Thornton CP (eds) (2014) Archaeometallurgy in global perspective. Methods and syntheses. Springer, New York/Heidelberg/Dordrecht/London 100. Roberts BW, Thornton CP (2014) Introduction. In: Roberts BW, Thornton CP (eds) Archaeometallurgy in global perspective. Methods and syntheses. Springer, New York/Heidelberg/Dordrecht/London, pp 1–9 101. Rovira Lloréns S (1996) Métodos analíticos aplicados a la investigación arqueometalúrgica. In: Calvo Peréz B, Bernárdez Pérez MJ, Guisado di Monti JC (eds) Arqueología e historia de la mineríay metalurgia. Escuela Técnica Superior de Ingenieros de Minas, Madrid, pp 27–41

24

Archaeometallurgy: A Discipline Between Past and Future

701

102. Cesareo R, Von Hase FW (1976) Analisi di Ori Etruschi del VII sec. a.C. con uno strumento portatile che impiega la tecnica della fluorescenza X eccitata da radioisotopi. In: Atti dei Convegni Lincei, XI Congresso Internazionale, Applicazione dei metodi nucleari nel campo delle opere d’arte, Rome-Venise 1973. Accademia dei Lincei, Roma, pp 259–296 103. Rovira S (1995) Estudio arqueometalurgìco del depósito de la Ría de Huelva. In: Ruiz-Gálvez Priego M (ed) Ritos de paso y puntos de paso. La Ría de Huelva en el mundo del Bronce Final Europeo. Servicio de Publicaciones, Universidad Complutense, Madrid, pp 33–57 104. Giardino C, Paternoster G (2019) Chemical-physical analyses by Energy Dispersive X-Ray Fluorescence (EDXRF). In: Giardino C (ed) Magan – the land of copper, prehistoric metallurgy of Oman. Archaeopress Publishing Ltd, Oxford, pp 139–163 105. Giardino C (2011) Matrici per vasellame metallico e altri residui di fusione di età ellenistica da Castro (Lecce). In: Giardino C (ed) Archeometallurgia: dalla conoscenza alla fruizione. Atti del Workshop (Cavallino-LE 2006), Beni Archeologici – Conoscenza e Tecnologie Quaderno 8. Edipuglia, Bari, pp 151–160 106. Giardino C, D’Annibale C (2017) Pizzica Pantanello (Metaponto): la più antica testimonianza di attività metallurgiche dall’Italia meridionale. AION - Annali di Archeologia e Storia Antica 21–22:47–63 107. France-Lanord A (1980) Ancient metals: structure and characteristics. Metaux Anciens: Structure et characteristiques. ICCROM, Roma 108. Rossi M, Casali F, Chirco P, Didonato T, Morigi MP, Nava E, Zanarini M, Morigi Govi C, Dore A (1999) 3D computed tomography: a tool for analysis and conservation of archaeological artifacts. In: Marabelli M, Parisi C (eds) 6th international conference on “Non-destructive testing and microanalysis for the diagnostics and conservation of the cultural and environmental heritage” (Rome 1999). I.C.R., AIPnD, Rome, pp 1595–1604 109. Matteini M, Moles A (1984) Scienza e restauro. Metodi di indagine. Nardini Editore, Firenze 110. Marabelli M (1994) The monument of Marcus Aurelius: research and conservation. In: Scott DA, Podany J, Considine BB (eds) Ancient & historic metals: conservation and scientific research. Getty Conservation Institute, Los Angeles, pp 1–19 111. Cagnetti V, Diana M, Ferretti M, Moioli P (1992) La Chimera di Arezzo: studio di alcuni aspetti metallurgico-strutturali mediante fluorescenza X e termografia as alta risoluzione. In: Marabelli M, Santopadre P (eds) 3a Conferenza Internazionale sulle Prove non Distruttive, Metodi microanalitici e indagini ambientali per lo studio e la conservazione delle opere d’arte – 3rd international conference on non-destructive testing, microanalytical methods and environment evaluation for study and conservation of works of art (Viterbo 1992). BetaGamma Editrice, Viterbo, pp 217–228 112. Scott DA (1991) Metallography and microstructure in ancient and historic metals. Getty Trust Publications (Getty Conservation Institute), Los Angeles 113. Sperl G (1980) Über die Typologie urzeitlicher, frühgeschichtlicher und mittelalterlicher Eisenhüttenschlacken. Studien zur Industrie-Archäologie VII. Österreichischen Akademie der Wissenshaften, Wien 114. Bachmann H-G (1982) The identification of slags from archaeological sites. Occasional paper 6. Institute of Archaeology, University of London, London 115. Gale NH, Stos-Gale ZA (1982) Bronze Age copper sources in the Mediterranean: a new approach. Science 216:11–19 116. Stos-Gale ZA, Gale NH (1982) The sources of Mycenaean silver and lead. J Field Archaeol 9: 467–495 117. Gale NH, Stos-Gale ZA (1987) Oxhide Ingots from Sardinia, Crete and Cyprus and the Bronze age copper trade: new scientific evidence. In: Balmuth MS (ed) Studies in Sardinian Archaeology III. Nuragic Sardinia and the Mycenean world, BAR international series 387. BAR Publishing, Oxford, pp 135–178 118. Stos-Gale ZA, Macdonald CF (1991) Sources of metals and trade in the Bronze Age Aegean. In: Gale NH (ed) Bronze age trade in the Mediterranean. Studies in Mediterranean archaeology XC. Astrom, Jonsered, pp 249–288

702

C. Giardino

119. Pernicka E (1992) Comments... III. Archaeometry 34(2):322–326 120. Stos-Gale ZA (1993) Isotopic analyses of ores, slags and artefacts: the contribution to archaeometallurgy. In: Francovich R (ed) Archeologia delle Attività Estrattive e Metallurgiche. All’Insegna del Giglio, Firenze, pp 593–627 121. Pernicka E (1995) Crisis or Catharsis in lead isotope analysis? J Mediterr Archaeol 8(1):59–64 122. Stos-Gale ZA, Gale NH, Houghton J, Speakman R (1995) Lead isotope data from the Isotrace laboratory, Oxford: Archaeometry data base 1, Ores from the Western Mediterranean. Archaeometry 37(2):407–415 123. Stos-Gale ZA (2000) Trade in metals in the Bronze Age Mediterranean: an overview of Lead Isotope data for provenance studies. In: Pare CFE (ed) Metals make the world go round. The supply and circulation of metals in Bronze Age Europe. Proceedings of a conference held at the University of Birmingham in June 1997. Oxbow Books, Oxford, pp 56–69 124. Dussubieux L, Golitko M, Gratuze B (eds) (2016) Recent advances in laser ablation ICPMS for archaeology. Springer, Heidelberg/Berlin

Part IV Applied Geophysics

Ground Penetrating Radar System: Principles

25

Mezgeen Rasol, Vega Pe´rez-Gracia, Francisco M. Fernandes, Jorge C. Pais, Sonia Santos-Assunc¸ao, and James S. Roberts

Contents 25.1 25.2 25.3

25.4

Ground Penetrating Radar Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GPR System and Working Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Type of Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.1 Dipole Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.2 Horn Antenna (Air-Coupled Antenna) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.3 Vivaldi Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.4 Spiral Antennas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.5 Antenna Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GPR Applications in Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

706 708 714 715 715 716 716 716 717

M. Rasol (*) Laboratoire Expérimentation et Modélisation du Génie Civil Urbain (EMGCU), Université Gustave Eiffel, Champs-sur-Marne, France e-mail: [email protected] V. Pérez-Gracia RMEE Department, EEBE School, Universidad Politécnica de Cataluña, Barcelona, Spain e-mail: [email protected] F. M. Fernandes University Lusíada – Norte, Vila Nova de Famalicão, Portugal e-mail: [email protected] J. C. Pais Department of Civil Engineering, University of Minho, Guimarães, Portugal e-mail: [email protected] S. Santos-Assunçao Department of Land Surveying and Geo-Informatics (LSGI), Hong Kong Polytechnic University, Kowloon, Hong Kong e-mail: [email protected] J. S. Roberts Met Consultancy Group, Leeds, UK e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_25

705

706

M. Rasol et al.

25.5

720 720 721 722 725 727 727 729 732 733

GPR Surveys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5.1 GPR Suitability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5.2 Antenna Frequency Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6 GPR Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7 Numerical Modelling of GPR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.8 Case Studies of GPR in Cultural Heritage Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.8.1 A Case Study from Spain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.8.2 Case Studies from Portugal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

Ground penetrating radar is a geophysical survey method widely applied to the assessment and monitoring of cultural heritage buildings. It is commonly used as a method of structural evaluation because it is nondestructive and noninvasive. This chapter describes the historical development of the method and explores the fundamentals and theory of ground penetrating radar systems and the properties of electromagnetic waves. Furthermore, it discusses some of the main applications and explains the procedure for data processing. Finally, it presents several case studies in the cultural heritage of Portugal and Spain.

25.1

Ground Penetrating Radar Development

Ground penetrating radar (GPR) involves the transmission and reception of electromagnetic waves. While the development of this technique has been taking place for over a century, marked improvements have led to widespread use during the last two decades. These improvements have resulted in recognition of the method as an accurate and efficient shallow test, with many civil engineering applications. One of the first uses of electromagnetic waves (EM) for remote sensing occurred in 1904 when Christian Hülsmeyer developed the Telemobiloscope. This device used EM fields to detect distant metallic objects [1–4], as shown in Fig. 25.1. Later, Heinrich Löwy and Gotthelf Leimbach applied EM fields to the detection of salt deposits and other mining applications [2, 3]. During the First World War, RADAR, the acronym for radio detection and ranging, was developed as the first reconnaissance and telecommunications equipment [5–7]. The method of using radar pulses, developed in 1926 by Hülsenbeck, was the first system to allow the detection of buried objects by identifying reflection events [8–10]. After this demonstration, the use of pulsed radar waves grows significantly due to its capacity for probing to considerable depths in ice [11, 12], desert sand, and rock [13, 14]. El-Said obtained the depth of the water table in 1956 by detecting signals reflected from the water table [15]. By having the transmitter and receiver a known distance apart, El-Said was able to measure the time delay of the received wave reflected from the water table in comparison to the direct wave between transmitter and receiver. This measurement allowed him

25

Ground Penetrating Radar System: Principles

707

Fig. 25.1 The Telemobiloscope system invented by Christian Hülsmeyer [4]

to calculate the depth of the water table, using procedures very similar to those used today [16]. Waite and Schmidt, in 1961, reported on the use of radar methods to image landmasses below polar ice, using a system deployed from an aircraft flying above the Greenland ice sheet [17]. Shortly after, in 1964, Walford published experiments demonstrating the ability of the technology and the accuracy achievable by “radio echo sounders” when used to investigate ice sheets in the polar regions [18]. Following the success of the method over ice, attention turned to investigating several other geological materials with similar properties [19]. In the 1970s, Cook [20, 21] and Roe and Ellerbruch [22] explored applications of the technique to investigate rock formations and coal deposits. Cook used the method in coal mines because of the low dielectric loss of the materials [23]. Holser et al., Theirbach, and Unterberger surveyed underground salt deposits [19, 24–26] as these are also relatively lossless.

708

M. Rasol et al.

This period also saw radar taken to the Moon. The Apollo 17 program using a pulsed radar sounder, similar to that used by Walford to investigate ice sheets, to take measurements from the lunar orbiter [19]. Sounding experiments were also carried out on the Moon itself using a receiver mounted on the lunar rover and a dipole antenna laid out across the surface [27, 28]. The use of GPR increased across a wide range of disciplines following these initial developments and documented use cases, with both science and industry recognizing the potential of the technique. Consequently, the number and variety of applications expanded significantly. The commercial production of GPR systems also started in the early 1970s [19]. Table 25.1 presents an overview of GPR development from 1974 to 1998. Because of this sustained development, GPR is now a standard nondestructive testing methodology in numerous fields. Peter Annan, the founder of Sensors and Software Inc., identified several points of note in his 2003 review [19], and these are even more true today after over 15 years of further development and refinement. We can see that: 1. GPR is a mature geophysical survey method with a well-developed theoretical basis in academic literature. 2. Instruments used for GPR prospection are of high quality and can be relied upon to produce consistent results. 3. Advanced data processing techniques have become common elements of readily available analytical software, allowing sophisticated analysis of GPR datasets using off-the-shelf products. 4. It is easy to derive complex visualizations of data from these advanced processing workflows and present them in a variety of formats, leading to improved interpretation of datasets. 5. Dedicated GPR systems with simple interfaces designed explicitly for restricted applications are allowing the widespread application of GPR by the non-specialist users.

25.2

GPR System and Working Principles

In general, most commercially available GPR systems consist of one or more antennas, one transmitter, and one or more receivers, connected to a control unit. This unit is a computer that controls the signal generation and the transmission and reception of wideband EM pulses. Additional accessories often used are the survey wheel or GPS, which triggers the system, and a display with a recording system [50], which may be integrated with the control unit. Figure 25.2 shows the different parts of a typical GPR system. Depending on the type and number of antennas, the target characteristics, and survey objectives, several types of field deployment are available. Figure 25.3 illustrates a simple schematic of a typical GPR system used in reflection deployment mode. The transmitting antenna releases EM waves into the ground, or structure, and the receiving antenna collects the reflections from underground elements. The

25

Ground Penetrating Radar System: Principles

709

Table 25.1 Overview of GPR development, from 1974 to 1998 Period 1970

Authors and institutions Morey, R Geophysical Survey System Inc. (USA)

1972

Ward et al. NASA

1973

Morey, Geophysical Survey System Inc.

1976

Annan and Davis

1976

Watts and England

1978

Dolphin et al., Stanford Research Institute

1980

Owen

1981

Coon et al

1981

A-Cubed Inc. (Canada) OYO Corporation (Japan) Ulriksen

1982

1984

Benson et al. Benson & Pasley

Work of note Rex Morey founded Geophysical Survey Systems Inc., which designed the first commercial ground penetrating radar (GPR) system used for service work in the USA. The company is still designing and selling GPR systems today Radar experiments are conducted on the Moon and from low lunar orbit GSSI changes company direction to focus on the development, manufacture, and sale of GPR systems Reported on the promise of GPR for investigating permafrost environments, driven by proposals for oil pipelines to be built linking arctic regions to consumer markets in the south Published paper on the effects of scattering in glaciers and identified need for lowerfrequency antennas in temperate regions due to absorption losses driven by higher ice temperatures GPR used in support of exploration and excavations at Victoria Peak Mexico. Demonstrated ability to identify voids and other features, and achieved depth soundings up to 400 ft Used borehole radar methods to identify tunneling. Investigations carried out on behalf of the US military This work investigated achievable penetration distances, propagation velocities, and attenuation at a range of frequencies and demonstrated the effectiveness of short pulse radar for investigating coal seams in advance of mining operations Rival companies to GSSI start development of GPR systems Investigated the applications of impulse radar to civil engineering applications, with work forming a foundation for future developments in applications such as pavement investigations and utility mapping Commercial drivers in the form of environmental initiatives to investigate contaminated land and hydrogeological applications resulted in further developments in GPR as a tool for subsurface mapping

Reference [9]

[19, 27] [9, 29]

[30]

[31]

[32]

[19, 33]

[34]

[19]

[35]

[9, 10]

(continued)

710

M. Rasol et al.

Table 25.1 (continued) Period 1987

1988

1990

1992

1993 1994

1995 1996

1999

2000 to present

Authors and institutions Olsson et al., Swedish Geological Survey

Work of note Reference Developing borehole GPR applications for [36] monitoring waste disposal in nuclear applications Annan P. Sensors & Sensors and Software company founded by [19] Software Inc. (Canada) Peter Annan to commercialize the PulseEKKO technology first developed by A-Cubed Inc. This decade saw an acceleration in the take-up and development of GPR, throughout industry and academia. Mala Geosciences formed in Sweden and ERA Technology in the UK began to develop ground penetrating radar systems [19]. The advancement in computing technology and power also allows for rapid development in the realm of GPR processing. Annan (2003) details several key advancements and publications in his brief review of GPR development [19], some of these are listed below: Fisher et al. Improving multi-fold data acquisition by the [37, 38] application of seismic data processing techniques to GPR data Maijala Application of seismic data processing [39] methods to GPR data Gerlitz et al Using digital data processing to improve the [40] resolution of near-surface targets Dean Goodman Advancements in data simulation and [41] processing for archaeological applications Brewster and Annan Using GPR in environmental applications, [42] specifically with uses for ground remediation Zeng et al. Using 2D numerical simulation to model [43] Cai and McMechan GPR data [44] Roberts & Daniels Using 3D Numerical simulation to analyze [45] polarization effects within GPR datasets Redman et al. Documenting improvements in the use of [46] borehole GPR for environmental applications through a presentation of several case studies Jol, H. Using GPR to investigate geological [47] stratigraphy and delineating sedimentary facies to investigate internal structural features of coastal barriers Holliger & Bergmann Designing and modelling 3D numerical [48, 49] et al. simulation using computer algorithms, tested against experimental observations The use of GPR has continued to grow exponentially, and it is now a standard tool used for surveying and prospecting within a broad range of disciplines and applications. Theoretical and computing approaches also receive considerable attention, with data processing methods continually being developed by different authors using methods for validating laboratory data through numerical modelling, combined with real-world testing. For an up-to-date review, Utsi (2020) has recently published a general non-technical introduction to the subject, with appendices that detail recent notable publications and listings of relevant conference proceedings and journals pertinent to GPR studies [49]

25

Ground Penetrating Radar System: Principles

711

Fig. 25.2 A typical GPR system and its main components. (Adapted from [100])

control unit is the component responsible for generating multiple single high energy pulses to the antenna, which are then transmitted through the study medium [51]. How a GPR system detects objects in the underground is dependent on the center frequency of the signal generated, the characteristics of the antenna, and the depth of the target. For correct interpretation, it is also essential to understand how the electromagnetic wave propagates and scatters throughout the medium. The electromagnetic wave, or radio wave, propagates in air at the speed of light, i.e., around 30 cm/ns. When the radio wave enters an appropriate medium, the velocity decreases according to its dielectric value. It is important to consider the amount of moisture present in the medium, as the value of the dielectric constant increases with moisture and, hence, the velocity decreases. This, in turn, affects the depth to which the signal can propagate. Normally, a single measurement corresponds to the time it takes for the signal to be transmitted by the antenna, reflected at the target and detected back at the receiver. This path is measured in nanoseconds and designated the two-way travel time; it relates to the depth and velocity of the radio wave, as per expression (25.1). The propagation depth is also affected by the signal’s attenuation – higher frequencies attenuate faster than low frequencies. v¼

2d t

ð25:1Þ

where v is the wave propagation velocity (cm/ns); t is the two-way travel time, measured in nanoseconds (ns); and d is the depth (cm).

712

M. Rasol et al.

Fig. 25.3 GPR system schematic. (Adapted from [55])

Apart from depth, the resolution, or the ability to discern closely placed objects, is fundamental. The resolution of a GPR system is determined by the wavelength and aperture of the signal. Wavelength is inversely related to frequency, which means that higher frequencies result in smaller wavelengths and lower frequencies result in larger wavelengths. To be able to distinguish the reflection of two different objects buried at depth, the vertical resolution is generally considered to be half of the signal’s wavelength [52, 53]. This means that if two targets are within half a wavelength of each other, they are not resolved. Additionally, to further distinguish distinct targets at the same depth, Snell’s law states that all objects within the area of the first Fresnel zone are detected as a single object. The area of the first Fresnel zone, the radius r is given by Eq. 25.2 [54, 55], is directly related to the signal’s wavelength (λ) and depth (z). The area of the first Fresnel zone is related to the antenna’s frequency wavelength such that the longer the wavelength, the larger the Fresnel zone. Therefore, high-frequency antennas are characterized by smaller Fresnel areas, and, consequently, higher spatial resolution

25

Ground Penetrating Radar System: Principles

713

due to their better capacity to distinguish closely positioned objects, when compared to lower-frequency antennas. rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi λ2 λz r¼ þ 16 2

ð25:2Þ

In practical terms, Fig. 25.4 presents a GPR antenna on two different targets (X is tree roots and Y is an embedded pipe) placed in different zones. Target Y is detectable because it is located inside of the Fresnel zone. On the other hand, the target Y is invisible as it is located outside of the elliptical cone of GPR antenna. Some amount of the transmitted wave propagates off-axis and outside of the elliptical cone of the emitted signal (not on the line-of-sight path between transmitter and receiver inside the elliptical cone). This can then deflect off of objects and then reflect back to the receiver. The path of the antenna could be described using the approach of a cone in the transmission of the energy (Fig. 25.4), being D the distance between the antenna and the target, A the radius of the path, εr the dielectric constant of the medium, and λ the wavelength of the wave. Resolution, as defined in Eq. 25.2 and Fig. 25.4, is therefore based on the ability to distinguish two close objects at the same depth. As the antenna path narrows, the possibility of detecting multiple objects as separated anomalies in the B-scan increases and, as a consequence, so does the resolution of the antenna.

Fig. 25.4 Elliptical cone of the GPR penetration modified from Conyers (2016) [55, 56]

714

M. Rasol et al.

However, a narrower antenna path does not always result in increased detection capabilities. In some cases, single small objects can be more easily detected if the number of scans reflected from their surface increases. Therefore, the possibility to detect these objects can be higher when the number of scans reflected from the target is increased, which can result from a system having a wider antenna path [55, 57]. Furthermore, the type of antenna is essential and influences the type of investigation and field deployment that can be carried out. Each antenna type has limitations in terms of where they can be used (shielded vs unshielded antennas), manageability (size of the antenna is generally related with wavelength), survey speed, and a host of other factors. Typically, commercial GPR systems consist of one or several antennas that are either the monostatic type, where the transmitter and receiver are located in a case, usually shielded, and with a fixed distance between them, or a bistatic antenna, where the transmitter and receiver are separated. While a fixed distance antenna is generally used for typical reflection data acquisition, bistatic antennas can be used for a range of other survey types. These include including common mid-point analysis (CMP), wide-angle reflection and refraction (WARR) surveys, and transillumination (tomography) measurements.

25.3

Type of Antennas

Antennas can be classified in different ways. One of them distinguishes between shielded and unshielded antennas. A shielded antenna, also known as a beam or directional antenna, is designed to transmit and receive signals in a particular direction, and, as a consequence, the highest part of the energy is radiated in a specific main beam. Usually, these antennas transmit and receive most of their energy through one of their surfaces. Unshielded antennas are omnidirectional devices, and energy is radiated and received equally in all directions. Another classification distinguishes between ground-coupled and air-coupled antennas. Ground-coupled antennas require contact with the surface of the medium for optimal operation. Air-coupled antennas are designed to be placed at about 40 to 50 cm away from the surface during operation. Ground-coupled antennas usually provide higher penetration depth than air-coupled antennas at the same frequency. However, air-coupled devices usually have a narrower beam path and allow higher speeds of data acquisition. Another mode of classification focuses on the specific properties of the radiating antenna used in the system. Both ground-coupled and air-coupled systems can be constructed using a variety of antenna types. Two types of GPR antennas have been traditionally used – these are classified as dispersive and nondispersive antennas. Some examples of the different types are as follows [58]: • Dispersive – Exponential slot, Vivaldi and spiral (logarithmic, equiangular, and Archimedean)

25

Ground Penetrating Radar System: Principles

715

• Nondispersive – Bow tie, TEM horn, biconical, resistive lumped element loaded, resistive continuously loaded The specific requirements of each application can define the best type of antenna for obtaining precise results from the survey. However, it must also be recognized that in many situations, it is the availability of any GPR system that may be the controlling factor on what is used for data collection. Finally, data can be acquired with a single antenna or with an array of antennas. An array is a group of antennas arranged and powered to obtain a pre-defined radiation pattern. Some of these typologies are described below.

25.3.1 Dipole Antennas Generally, a dipole antenna is a type of ground-coupled antenna that presents a linear polarization and a relatively limited bandwidth. Barring loading or distributed loading techniques are utilized to increase the bandwidth at the outlay of radiation efficiency [58]. This type of antenna is cheap and easy to implement. This type of ground-coupled antenna works in a broad range of central frequencies. The usual range is between 80 and 1500 MHz, and the proposed depth of the surveys can range from 20 to 30 m to just a few centimeters. When using this antenna, the transect distance is often the most significant issue because as surface conditions change so does the quality of ground coupling during data acquisition. Its advantage is the higher signal penetration than with air-coupled system because of surface coupling and decreased ringing. On the other hand, the vertical resolution of a ground-coupled antenna is better. Several commercial companies use this kind, including GSSI (USA), IDS (Italy), MALA (Sweden), Penetradar (USA), Sensors and Software (Canada), 3D-Radar (Sweden), Impulse Radar (Sweden), UTSI Electronics (UK), and Proceq (Switzerland) [59].

25.3.2 Horn Antenna (Air-Coupled Antenna) Horn antennas are a type of air-coupled antenna. They can be a transverse electromagnetic (TEM) horn or a conventional horn. The TEM horn is quite often used in GPR systems because of its excellent characteristics in the time domain. It consists of two metallic plates diverging from a feeding point. This type of antenna is achieved by resistive loading. A TEM horn antenna is nondispersive, directive, and ultra-wideband (UWB) [60]. This kind of antenna is widely used in road pavement investigations [59] and to detect layer thickness and layer dielectrics [61]. They have the advantage of being able to be deployed at high speeds and so allow investigations to be carried out without disrupting normal traffic flow [59]. These air-coupled antennas are an impulse radar system. They usually operate at center frequencies between 400 MHz and 2.5 GHz, and the depth penetration can be between 0.5 and 0.9 m

716

M. Rasol et al.

[61]. During data collection, the horn antenna is suspended between 0.3 and 0.5 m above the surface being measured. This kind of air-coupled system is most widely based on the transverse electromagnetic (TEM) horn antenna. But other types have been used for road surveys such as hemispherical butterfly dipole (HBD) antenna. The main advantage of the air-coupled antenna is that the coupling does not change when surface properties change. Furthermore, in the absence of interference from traffic, the speed of data collection can be up to 100 km/h at full speed. Several commercial companies manufacture air-coupled antenna including GSSI, Penetradar, Pulse Radar, and Wavebounce [59, 60].

25.3.3 Vivaldi Antennas A Vivaldi antenna is a type of end-fire tapered slot antenna. This type includes both linearly and exponentially tapered solutions; however, it is only the exponential solution that is known as the Vivaldi type [16]. Vivaldi antennas are often used in GPR systems and were first introduced by Gibson in 1979. The main advantages of this type of antennas are the wide frequency band, medium directivity, and low sidelobes [16, 62]. From this system, the result of characterization achieved is constant [60, 63]. Researchers have experimented with using metamaterials to improve the directivity of Vivaldi antennas [60, 64].

25.3.4 Spiral Antennas This antenna takes the form of a spiral shape and when constructed can theoretically be considered a frequency-independent device; however, the bandwidth is finite as it is related to the antenna arm length and the width of the inner gap [16]. A spiral antenna has the capacity for circular polarization which makes it an ideal candidate as a GPR antenna for detecting circular objects and because of this has been the focus of research for system development with the aim of landmine detection and clearance [65, 66].

25.3.5 Antenna Arrays Arrays of antennae can be used in GPR systems to make the data collection process faster and to extend the investigation of the field area per time unit [59]. The recent developments of GPR array antenna, vehicle towed GPR, and multi-channel antennas provide an opportunity to explore very dense radar datasets across a wide frequency bandwidth. They can provide both a detailed resolution of near-surface targets and an improved depth of penetration in site investigations.

25

Ground Penetrating Radar System: Principles

717

The advantages of an antenna array can be seen in several applications: mine detection [67], archaeological prospection [65], bridge inspection [68], utility mapping [69], ad crack detection [70–73] in addition to a host of other civil engineering and geoscience applications. There are two main design approaches to GPR antenna arrays that are usually taken. The first approach is used to visualize multi-channel radar composed with multiple single-channel radar units, and the second approach is to synthesize the whole system of antennae. Some advantages of the antenna arrays are the following: 1. Improved imaging capabilities 2. Faster area coverage 3. Provides dense 3D datasets

25.4

GPR Applications in Cultural Heritage

GPR is widely used in the industry and academia across a wide variety of sectors, described thoroughly in Table 25.3. The most common applications in cultural heritage deal with archaeological and building surveys, and forensics, as described below. Archaeologists and forensic investigators have used GPR to map archaeological sites and buried human remains [19]. The crossover arises from the fact that often buried human remains cannot immediately be identified as being of archaeological origin and vice versa. James Mellett used GPR in 1990 to examine the upper 2-m-depth earth surface [74] identifying the sensitivity of GPR velocities to saturation. In 1992, using GPR on an American Indian historical cemetery site, potter’s fields, he detected human remains related to a missing person case. That’s why this survey involved law enforcement as well as archaeologists [66]. At this time, many investigations in historical cemeteries (grave sites) were taking place, as described in Table 25.2. All scans shown in Table 25.2 were carried out using Geophysical Survey System Inc. Model SIR-3, 3102–500 MHz antenna. Two points are vital to be known by GPR users. The following factors can allow burial sites to be found using GPR for many years after internment [66]: (a) Electrical properties of disturbed subsoil, coffin furniture (e.g., handles, plaques), and artifacts (i.e., clothing, decorations) can render graves detectable [75]. (b) Lain bones in the soil can be detectable even after many years of burial due to: (i) Chemical composition of bone (i.e., dry bone weight consist of 20% collagen and 80% hydroxyapatite) [76]. (ii) Bone structure, two phases material, consisting of crystalline material (i.e., Ca (OH)2 and 3Ca (PO4)2), organic materials (collagen), and groundmass.

718

M. Rasol et al.

Table 25.2 First GPR scans done in criminal and archaeological cases Scans no. 1st scan 2nd scan 3rd scan

4th scan 5th scan 6th scan

Scan date March 16, 1989 August 17, 1989 March 24, 1990

Cemeteries location Baptist Cemetery, Jamaica, Vermont Gethsemane Cemetery, Little Ferry, NJ Missing Persons Case Eastern United States

November 17, 1990 June 19, 1991 October 10, 1991

Salwen Archaeological Site, Bergen County, NJ Potter’s Field, Baltimore, Maryland Potter’s Field, Westchester County, New York

Carried out by Rolfe Johnson and Joseph Miramontes Dr. Joan Geismar and Ruth Van Wagner Tom Fenner, Pete Petrone, Wayne Saunders, Jordan Watts, End Naylor, Wayne Murphy, and Mike Greco Eugene Boesch and Arnold Pickman Sam Bowerman

Referred [66]

Frederick C. Drummond

Barone et al., in 2004, documented investigations undertaken at St. Agata La Vetere Church in Catania, Italy, which included GPR surveys among a variety of other geophysical and novel investigative techniques. The church and adjoining monastery had been destroyed because of an earthquake in 1693 [77]. Techniques used included optical microscopy (OM), scanning electron microscopy (SEM), X-ray diffraction (XRD), differential scanning calorimetry (DSC), Fourier transform infrared spectroscopy (FTIR), and thermo-gravimetric (TG) methods to examine bricks, plasters, and mortar. Ground penetrating radar (GPR) was applied to detecting anomalies related to buried structures, and human remains both inside and outside the main buildings [75, 77]. Building condition assessment is an integral element of structural inspections. The data is often collected by visual assessment and is used to inform requirements for retrofitting or strengthening buildings. However, the use of GPR and other nondestructive techniques for structural inspection has widely increased in recent years [19, 59]. Nuno Barraca et al., in 2016, presented a case study demonstrating the use of GPR in several applications for building assessment (i.e., structure, construction technique, history age of structure, and materials). In this study (Art Deco Building, Portugal), they obtained several findings: first, the overall quality of the results was increased due to the use of nondestructive methods; second, the investigation was carried out with no additional damage caused to the existing building; and third, the building was an archaeological or heritage structure due to its architectural features and the age of the building [78]. Finally, they highlighted the importance of choosing the correct antenna frequencies and the importance of careful data processing of the examined structure (Table 25.3).

25

Ground Penetrating Radar System: Principles

719

Table 25.3 Actual GPR applications Field Civil engineering

Mining industry

Hydrology

Architecture

Cultural heritage

Archaeology

Agronomy

Forensic medicine

Glaciology

Applications Ground investigations to inform foundation design Crack detection in structures Assessment of oxidation and water content in reinforced concrete structures Soil identification under rigid frames Identification of cavities in calcareous soils Location of buried structures Assessment of bridge decks and rebar Investigation of road pavements Airport assessments Railway assessments Seismic zonation and assessment Void detection Detection of mineral deposits Determination of water levels in gas or oil reservoirs Location of conduits, manholes, and chambers Determination of water level Identification of aquifers and tree roots Study of diffusion of pollutants Determination of connections between aquifers Assessment of structural damage Locating utilities Detection of hidden structures Quality control of maintenance and repairs Assessment of structural integrity Investigating known and unknown construction features Determining the presence of corrosion Location of architectural remains Detection of cut and fill features Identification of graves Mapping buried structures Control of soil maps Determining depth to bedrock Identifying water level Determining presence/absence/extents of contamination Determination of surface moisture content Locating of buried bodies Detection of hideouts and illegal tunnels Identifying burial sites Depth estimation of ice sheets Determining underlying topography Geomorphological study of glaciers (continued)

720

M. Rasol et al.

Table 25.3 (continued) Field Armed forces Marine scientific research

25.5

Applications Identification of landmines and hideouts Locating buried caches Determination of marine geomorphology Identification of sedimentary structures

GPR Surveys

If a GPR field survey is to be carried out, it should be planned and designed correctly. The conditions for data acquisition in the field must be carefully assessed to optimize the chances of collecting data that can address the aims of the investigation. The expected results after data processing and interpretation should be thought about in advance of the survey, as this can affect how the field survey is undertaken. Several parameters must be considered during the designing of the field survey. Survey planning should take account of the radar frequency to be used, ground conditions, the type of deployment (cart-based, vehicle-mounted, towed), site access, the time window to be used, soil type, the sampling interval and line spacing, temporal sampling requirements, the nature of the target (or targets), weather, site hazards and risks, and antenna separation (especially in the case of special measurements, e.g., CMP, WARR) and the general public.

25.5.1 GPR Suitability The suitability of the GPR for the intended application must be taken into consideration before any field survey is conducted [79]. Otherwise, it risks wasting time and cost. In general, the field area has to be inspected, and objects that could prevent adequate data acquisition or affect the GPR signals should be assessed and removed, where possible. Obstructions such as metallic objects, vehicles, vegetation, livestock, and underground infrastructure, [53, 80] can all impede the successful conclusion of a GPR survey. The problem of underground infrastructure is especially relevant in urban areas and particularly when the buried infrastructure is not the focus of the investigation. One must also take into consideration the material properties for GPR suitability. Most construction materials are suitable for a GPR survey. Soil, due to its heterogeneity, is more difficult to assess. However, the USA has a map of soil suitability for GPR [81, 82], which identifies where the soil can provide a good response, or no survey can be carried out. Several factors were underlined by Robinson et al. [4] to be considered before choosing GPR for a specific application:

25

Ground Penetrating Radar System: Principles

721

• The depth of penetration in specific applications could be one of the critical limitations of the GPR system and depends on the use of an appropriate frequency that can reach the required depth for the field survey to have a significant result. • The electrical properties (i.e., dielectric constant and electrical conductivity) of the soil and ground conditions where the survey is to be conducted. When the subsurface comprises different layers, and the radar pulse reaches the interface between two layers, some energy is reflected and some transmitted. The reflected energy is expressed as a reflection coefficient (R) in Eq. 25.3. The variation of the R should be in a range of +1 and 1. pffiffiffiffiffi pffiffiffiffiffi v2  v1 R ¼ pffiffiffiffiffi pffiffiffiffiffi v2 þ v1

ð25:3Þ

where v1, v2 are the velocities of layer 1 and layer 2 of the subsurface.

25.5.2 Antenna Frequency Selection The selection of antenna frequency is based on the survey application and the required resolution and expected depth of the target of interest. High-frequency waves can detect and resolve small features but to relatively shallow depths. On the other hand, the low-frequency waves exhibit higher depth penetration, at the expense of a lower resolution, preventing detection of small features. The selection of an appropriate antenna and frequency depends on several factors that affect the accuracy level of the data collected from the field. These factors are the following [83]: (a) (b) (c) (d)

Spatial resolution Depth of penetration Clutter limitation Survey application

Spatial resolution and center frequency ratio to bandwidth to 1 can be considered for the constraint of center frequency fc as indicated in Formula 25.4, where Δz is the spatial separation (m) and K is the dielectric constant (relative permittivity) [84, 85]. f Rc >

75 pffiffiffiffi MHz Δz K

ð25:4Þ

The radar signal that is propagated to the heterogeneous medium is clutter; increasing the frequency causes the increase of the clutter and signal response. The depth of the media could correspond to the decreasing of the energy scattered by the clutter. This is indicated in the form below, where ΔL is the dimension of the clutter:

722

M. Rasol et al.

f Cc >

30 pffiffiffiffi MHz ΔL K

ð25:5Þ

The frequency required for penetration depth is based on the radar beam and allowed energy reflected to detect the buried target. The size of the target should be close to that of the Fresnel zone. Therefore, the receiver antenna should detect the reflected signal, and its frequency is expressed in Eq. 25.6, where D is the depth in meters, β is the radar beam to the target dimensions ratio, and in a GPR system D is presumed to be 4 as expressed in Eq. 25.7. pffiffiffiffi K1 MHz > D pffiffiffiffi 1200 K  1 D fc > MHz D fD c

vβ

ð25:6Þ ð25:7Þ

In order to validate the results, comparison of the frequency constraint computation is expressed as below: f Rc < f c < min



f cD , f Cc



ð25:8Þ

The compatibility of the spatial resolution with clutter size and penetration depth is based on the resolution frequency. For this reason, the resolution frequency should be superior to the other resolution. In Table 25.4, the frequency selection, according to field application, is presented. The expected depth of penetration as a function of the frequency is indicated in Table 25.5. The sampling frequency should be at least double the central frequency of the antenna.

25.6

GPR Data Processing

Raw data is the information directly obtained from the GPR during the field acquisition without any filtering. The data is therefore imported to a processing software to optimize the quality of the radar image to obtain the most information as possible, eliminating unwanted information. There are several commercially available software packages for data processing; some of them are general purpose while others are specific to a particular application or GPR system. Examples of independent software programs include GPRSlice, created by Dean Goodman, and ReflexW, created by Karl-Josef Sandmeier. Each of the leading GPR manufacturers provides a proprietary data processing solution, such as GSSI’s Radan package, GredHD by IDS, ObjectMapper from Mala, and EKKO_Project, by Sensors and Software, among others. The sequence to process filtering is variable and depends on many factors. The survey location and the medium being investigated play an essential role when carrying out a GPR survey. Dense cities are usually an immense source of noise, widely spread on the electromagnetic spectrum. Radio channels emit AM and FM

25

Ground Penetrating Radar System: Principles

723

Table 25.4 Various applications as a function of an appropriate center frequency range Applications Field of study Civil engineering

Mining industry

Hydrology

Architecture

Cultural heritage

Archaeology

Agronomy

Forensic medicine

Antenna central frequency (MHz) Ground investigations to inform foundation design Crack detection in structures Assessment of oxidation and water content in Reinforced Concrete Structures. Soil identification under rigid frames Identification of cavities in calcareous soils Location of buried structures Assessment of bridge decks and rebar Investigation of road pavements Airport assessments Railway assessments Seismic zonation and assessment Void detection Detection of mineral deposits Determination of water levels in gas or oil reservoirs Location of conduits, manholes, and chambers Determination of water level Identification of aquifers and tree roots Study of diffusion of pollutants Determination of connections between aquifers Assessment of structural damage Locating utilities Detection of hidden structures Quality control of maintenance and repairs Assessment of structural integrity Investigating known and unknown construction features Determining the presence of corrosion Location of architectural remains Detection of cut and fill features Identification of graves Mapping buried structures Control of soil maps Determining depth to bedrock Identifying water level Determining presence/absence/extents of contamination Determination of surface moisture content Locating of buried bodies Detection of hideouts and illegal tunnels Identifying burial sites

500–2600 1600–2600 1000–5000 100–500 10–500 100–500 800–3200 800–3200 800–3200 800–3200 100–250 800–3200 80% of domestic cooking pottery assemblages worldwide [14, 15]. The characterization of the compoundspecific molecules from the complex absorbed organic mixtures has become a wellestablished methodology and a standard part of a well-designed archaeological research program, [14] which can contribute to the better understanding of the material culture and allows to challenge and reassess traditional models [15] on ceramic function and vessels use, diet, culinary and eating habits, economic activities, trade, and trade networks. On the application of chemical analytical techniques for the detection and identification of the contents of archaeological ceramic material (i.e., pottery vessels and amphorae), a consistent amount of literature is already available, starting from the 1970s [16–20]. The main discriminant is determined by the kind of organic residues to be analyzed in terms of visible adherent or amorphous absorbed residues. Visible residues can be analyzed by means of optical microscopy like scanning electron microscopy (SEM) or spectroscopic methods, i.e., infrared (IR) spectroscopy and Raman spectroscopy [6, 21]. Visible residues are barely preserved and, if preserved, even more subjected to contamination from soil. [22] Recently, a proteomics approach on calcified deposits adhering to the inner surface of ceramic sherds from Çatalhöyük shown excellent results in the detection of the evidence of dairy products by integrating the analysis of the lipids absorbed within the pottery with the extraction of ancient proteins from the vessels [23]. The analysis of the microscopic organic residues makes use of several biomolecular methods and techniques [24]. According to literature, GC and GC-MS are the most suitable techniques for the analysis of absorbed amorphous organic residues in archaeological materials after solvent extraction according to different analytical protocols, [25] as it is able to separate (via gas chromatography) and characterize (via mass spectrometry) the individual components of complex mixtures of lipids entrapped within the porous mineral matrix of the ceramic fabric, where they are

37

Organic Residues Analysis (ORA) in Archaeology

1079

often preserved from microbiological degradation [15]. Nevertheless, the analysis of non-visible residues with biomolecular techniques is not straightforward, as they are complex mixtures of different compounds.

37.2

Sampling Procedures: Excavation, Post-excavation, Transport and Storage

When doing ORA in archaeological contexts, the most difficult decision arises already before the analysis starts since only few methods allow a non-destructive analysis. Therefore, samples must be sorted: Which samples should be fully conserved? Which methods for analysis are more suitable to obtain the information needed? Which convenient micro-invasive techniques do exist (see Sect. 37.3)? This rationalization is easier when dealing with large contexts of very similar finds [26]. Impacts by deliberate adulteration [27] or attempts of conservation of artifacts [28–31] on ORA are not discussed within this chapter, although they might have huge effects on the findings. A precondition of a successful analysis of the organic content of archaeological artifacts is a precise and thorough documentation as well as a proper treatment of the entire pathway of the find. It starts with a detailed description of the find’s excavation context, especially the find spot. This includes all relevant information like the available geochemical data (i.e., humidity of the surroundings (for effects, see, e.g., [26]) soil composition, climate, and climate changes possibly having an impact on the site) [14, 21, 32–34]. Other important points are possible sources for adulteration of the find, like signs for burns, [35–37] water courses (organic), content of earth layers deposited in the artifact’s surrounding, [38] and also exposure to sunlight (see, e.g., [39]) and air until analysis. For further reading on archaeological projectdesign including ORA from the beginning, we suggest a publication of Historic England on ORA best practice [14]. A proper treatment of the artifacts while performing the excavation is also important. If possible, the excavated artifacts should only be cleaned on the surface mechanically by clean tools. Samples of the soil surroundings are a plus when looking for possible contamination or chemical background. Nevertheless, comparative analysis of sediment and residues shows that the contamination of the surrounding soil is negligible [21, 38]. Also, a sterile working atmosphere in laboratories with access to high-purity solvents is often crucial to avoid contamination or removal of the original organic content. Wearing sterile powder-free gloves is required for protecting the sample any time. Not only does the skin contain lipids, but skin-care products do contain a lot of components hard to be removed later (e.g., PEGs). Another source of contamination is using improper packaging material (e.g., aluminum foils usually have oil containing separating agents on the surface, and plastic containers or reaction vessels might contain a high content of plasticizers, monomers, etc.). Once excavated, the samples should be protected as much from outside effects as possible. This includes

1080

S. Polla and A. Springer

protection from external dust or contamination by contact with the packing material itself and possibly avoiding access to humidity, air, or light as well.

37.3

Sample Treatment

The method of analysis affects not only sample preparation but also the artifact itself as well. There are promising attempts in developing non-invasive or microdestructive sampling techniques [40–43] or instrumentation [44–46]. Developments in the method and instrumental section not only offer more information in advance, for example, on the amount of organic residues in an artefact, [47] but also the amount of sample needed for the analysis is reduced. This is, for example, the case of radiocarbon dating nowadays, where only about 1‰ or few mg of the sample instead of several g are required after switching from decay counting methods to IR-MS (also see Sect. 37.6) [48]. Still, some artifacts are too big or too fragile to be moved, and some instruments have high requirements on environmental conditions, needs for supply, or training of the personnel. Then using a (micro)-invasive sampling technique cannot be avoided. However, often the place of sampling can be chosen in a fashion that does not affect the visual appearance of the artifact. Most spectroscopic techniques like UV/vis, Raman, and IR spectroscopy [40, 42, 46] allow a non-invasive or, more frequently, micro-destructive analysis with no or a very limited sample preparation (compared to around 1 g sample for extraction of lipids or wine markers from ceramics). Generally, these methods provide a fast and reliable overview over the sample composition (see Sect. 37.8 for further information) but have some drawbacks when a theory needs verification and specific marker molecules like sterols are targeted. Direct analysis with different techniques of mass spectrometry (MS, see Sect. 37.7) enhances the specificity of the information gathered. Also, they do not require much sample preparation. Nevertheless, complex samples usually result in spectra which are hard to interpret. When the compound class-specific spectroscopic methods or the direct analysis by mass spectrometry is not enough, separation techniques like gas chromatography (GC) and GC-MS (see Sect. 37.4) or liquid chromatography and LC-MS (see Sect. 37.5) are required. The demands on sample preparation for chromatographic methods are higher than for the methods described before. Both kinds of chromatography require that the sample is dissolved completely inside the solvent used before analysis, and more restrictions apply. Therefore, the compounds of interest must be extracted from the non-dissolving part. Depending on the compound classes and the following analysis, different protocols are used [49, 50] For LC-MS, the compounds should be polar to ionic. For GC-MS, the compounds should be non-polar and volatile. Separation, volatility, polarity, and ionization efficiencies can be altered by several derivatization strategies [49]. As an alternative to these methods, pyrolysis GC-MS (see Sect. 37.4) offers a method to create defined pyrolysis products from polymeric or other involatile samples, which then can be separated and analyzed by GC-MS.

37

Organic Residues Analysis (ORA) in Archaeology

1081

37.3.1 Sample Preparation, Extraction, and Derivatization In order to consider different aspects of sample preparation, a detailed description of one method described by Pecci et al. in 2013[1, 51] will follow. It is suitable for the analysis of wine markers and other small organic acids as well as small fatty acids. The original procedure is written in italics, and explanations or alternatives are added as standard text.

Sample Preparation In the first step, samples of ceramic vessels or other samples are cleaned using a scalpel or alternatively a drill (Fig. 37.1). In case of an alternative cleaning with solvents, the danger of irreversible dissolution and loss of compounds would be the result. We additionally use high purity nitrogen to blow away loose particles possibly causing contamination. In the second step, the samples are ground to obtain a fine powder. This increases the accessible surface for the following extraction. In case of visible deposits or different-colored layers, these deposits/layers were removed carefully and collected as an extra sample. Later, comparison of the results of the cleaned ceramics material with results of adhered material and a third set of material, loosely attached soil, can give insight into differences. This increases certainty of what was inside the analyzed material and not coming from outside contamination. Sample Extraction For sample extractions, about 500 mg of the sample is extracted with KOH (1 M, 3 mL) in water, in a sonicated bath at 70
C for 90 min. The KOH increases polarity of the target molecules, organic acids, and other polar and at least slightly acidic

Fig. 37.1 Sample preparation

1082

S. Polla and A. Springer

compounds by deprotonation. Sonification assists the dissolution process better than vigorously shaking the sample. Additionally, a long extraction time and an elevated temperature are used to speed up reactions like hydrolysis. These conditions lead to hydrolysis of triglycerides, phospholipids, and other esters, while more inert compounds like ethers and amides usually remain unchanged. The high temperature increases the danger of loss of volatiles, so working in vessels which can be closed very tightly is recommended. In the next step, the remaining solid is removed using centrifugation after cooling of the sample. Filtering, especially under such basic conditions, might lead to contamination of the sample with filter material, and some compounds might be retained and not completely eluted. The following step is crucial for success of the analysis. After recovery of the supernatant, it has to be acidified with approx. 15 drops of concentrated HCl. The pH should be controlled to be not higher than 1, since a too high pH leaves the goal molecules (somewhat acidic molecules) deprotonated and polar, while other compounds still are nonpolar. For example, two important biomarkers have low pKs values: tartaric acid has two pKs values (2.98 and 4.34), and oxalic acid has even lower pKs values (1.23 and 4.19) and thus are prone to be lost. Then, 3 mL of ethyl acetate is added to the acidified solution and mixed by vortexing for 2 min. Ethyl acetate is a nonpolar solvent and is not mixable with water. The more nonpolar a compound is, the better it will dissolve in the ester and the less remains in the acidified water phase. Vortexing speeds up the equilibration by efficiently mixing the two phases, increasing the interaction area. After vortexing, the phases are separated by centrifugation. Like vortexing for mixing, centrifugation speeds up the process and efficiently separates the phases. The supernatant is recovered and dried using a gentle stream of nitrogen. In this case, the organic phase is lighter than the water phase, in which all the salt from neutralizing the KOH is dissolved. Using a gentle stream of nitrogen allows efficient removal of solvent without loss of more volatile compounds, whereas using a rotavapor or similar device using heat and/or reduced pressure leads to significant loss. The procedure is repeated three times. Repetition is significantly more efficient than initially using 9 mL and doing the whole procedure once. For every extraction, the equilibrium between the phases is the same, but with less sample left in the water phase, in total more will go into the ester phase. We also observed loss of compounds when not performing the extraction step right after acidification. Many potential biomarkers undergo reactions in acidic conditions. The extraction does not only work for the small, organic acids but also is capable to extract some fatty acids. An extraction protocol for lipids (see, e.g., Charters et al. [12] and Mottram et al [52] or slightly modified [2]) enhances the chance to detect sterols, long-chain fatty acids, and alkanes, specific markers for animal or plant origin and important for waxes and resins. Nevertheless, using such a protocol will lead to a loss of all small and polar organic acids as well as of even longer-chain dicarboxylic acids like pimelic or suberic acid due to the low polarity of the first solvent used (2:1 chloroform:methanol).

37

Organic Residues Analysis (ORA) in Archaeology

1083

Derivatization, Dilution, and Measurement The extracts are derivatized by adding 25 mL of N,O-bis(trimethylsilyl) trifluoroacetamide (BSTFA) and heating at 70
C for 1 h. All solvent should be carefully removed before the reaction. Especially when some water is left in the sample by accident, the derivatization agent is hydrolyzed and not reactive anymore. Since the boiling point of BSTFA is significantly below 70
C, reaction vessels which are sealing efficiently should be used. The elevated temperature speeds up the reaction, and the long time is needed especially to derivatize less reactive functional groups like alcohols or to fully derivatize a molecule with more functional groups. For example, the important wine marker tartaric acid has four groups which must react, with the two alcohol groups being less reactive; otherwise retention times and observed mass spectra will not match and even worse split into several peaks in the chromatogram. BSTFA is leading to the trimethyl silyl adducts of OH and, if present, NH and SH groups. Similar silylation agents like N-Methyl-N-(trimethylsilyl)trifluoracetamid (MSTFA) can be used as well, and alternative methylation reactions are described as well, giving the more stable methyl esters and ethers using agents like a 14% w/v boron trifluoride–methanol complex [52] or trimethylsilyldiazomethane (TMSD) [53] but also more toxic and carcinogenic agents like diazomethane [54] and iodomethane (MeI). Methyl and TMS derivatives will have different retention times and peaks in the mass spectra. Then, 75 μL of hexane and 5 μL of a standard solution of dotriacontane (1 mg/ mL) are added. Of this solution, 1 μL is injected for analysis. After dilution with hexane, the sample can be analyzed with gas chromatography. In some cases, we did observe that not everything was dissolved, even after vortexing for a couple of minutes. Since this would be harmful for the instrument, the solid was removed by centrifugation. The supplement of dotriacontane adds a defined peak to the chromatogram (internal standard); a compound like those analyzed and close to the end of the chromatogram was chosen not interfering with marker compounds. Changes in intensity and/or retention times allow determination of issues with the instrumentation. A standard added to the sample before extraction would allow to follow the whole procedure, including hydrolysis/extraction.

37.4

Gas Chromatography (GC): Potential and Limits

Using a separation technique offering a very high peak capacity like gas chromatography (GC) offers unique chances in analysis of unknown compound mixtures, with basically one limit: the analytes have to be volatile enough to go into the gas phase without decomposition when heated to approx. 300
C. The volatility of polar compounds like organic acids or alcohols can be enhanced by derivatization. Often esterification/etherization using a silylating or methylating agent is performed before analysis. When the GC is coupled to a high-sensitivity multi-channel detector, this

1084

S. Polla and A. Springer

method is ideal for the analysis of trace amounts of complex mixtures. It is one of the most robust and reliable methods in analytical chemistry, while still affordable. Therefore, it is one of the most widespread methods in the world, with thousands of possible applications. It allows the comparison of known compounds with signals in the (archaeological) samples providing high certainty with a very low rate of false positives.

37.4.1 Gas Chromatography with Non-MS Detectors When not coupled to a mass spectrometer, detectors like FID (flame ionization detector, sensitive for hydrocarbons), TCD (thermal conductivity detector, universal use), or ECD (electron capture detector, for analytes with electron withdrawing functional groups like such as carbonyls) [55] among others are used. Since its development in the late 1940s, hundreds of applications have been developed. GC remains one of the standard methods for qualitative and quantitative analysis, since it offers a high sensitivity as well as a high resolution [49, 56]. Signals derived from the detectors mentioned above are not containing information about the molecular structures. Therefore, the analysis of unknowns and very complex mixtures remains an issue. Sample preparation, especially the choice of the extraction protocol, can limit the number of possible analytes at a fixed retention time. Subsequently, comparison with known standards and/or comparison of retention times on columns with different selectivities [49, 56, 57] is possible. But when it comes to unknowns, the method comes to its limits. Nevertheless, the numerous existing methods in analytical, food, or pharmaceutical chemistry as well as clinical and forensic science can be easily transferred to questions in the field of ORA in archaeology. Among others, examples of successfully using a GC-FID for the analysis of ruminant fats [58] or beeswax [59] are described in literature.

37.4.2 Gas Chromatography–Mass Spectrometry (GC-MS) GC-MS (gas chromatography – mass spectrometry) is one of the standard methods in organic residue analysis [25, 28, 60–63]. This is because it is one of the most costeffective and at the same time most versatile and sensitive methods. With EI as ionization source, the derived mass spectra are very similar for different instrument types [64]. This allows the comparison with mass spectra deposited in libraries like NIST [65] or Wiley, [66] generating high probabilities in the assignment of previously unknowns or unexpected compounds [67]. A more recent trend is the theoretical prediction of EI mass spectra, [68, 69] enhancing the referable compounds even more. A GC-MS consists of several building blocks, with a short description and their principal mode of operation described below (see Fig. 37.2).

37

Organic Residues Analysis (ORA) in Archaeology

1085

Fig. 37.2 Schematic drawing of a GC-MS instrument

Mobile Phase and Gas Flow Generation A mobile-phase reservoir, typically a pressurized gas container with high purity gases like He or H2, less frequently N2 or compressed air (due to drops in performance but better availability), and rarely other gases, is connected to a regulator generating a constant gas flow by precisely controlling and adjusting the pressure or the gas flow [49, 56]. In many actual GCs, the flow can be programmed, even though this is rarely performed within a separation. Injector This set-up is followed by the injector unit. In ORA, the most common injector type is a split/splitless injector which is typically heated. The sample solution is vaporized after the injection performed with a microliter syringe, limiting the analytes to volatile, thermally stable compounds [49]. A special method often used in ORA is py-GC (pyrolysis-GC). py-GC typically is used for nonvolatile analytes such as resins containing polymeric compounds. The sample is heated very fast to high temperatures (usually between 400
C and 4000
C), [63] leading to (reproducible) thermal degradation products [49, 70]. In any case, the vaporized analyte mixture is transported into the separation column by the gas flow. GC Oven and Separation Column (Stationary Phase) The analytes are carried with the gas flow through a GC separation column which is located inside a programmable oven. The separation is performed due to different volatilities of the analytes and differing interaction with the GC column. The latter is based on solubilities of the analytes inside the thin film inside the GC capillary column (absorption) or adsorption on the surface of the columns’ inner surface (adsorption). If the mass transfer as well as the equilibrium constants of two analytes are different, the compounds can be separated. The repeated equilibration can be altered by the choice of the separation column. The choice of the carrier gas plays a minor but sometimes significant role. The stationary phase usually consists

1086

S. Polla and A. Springer

of a thin polysiloxane film with the end groups defining the polarity [49, 56, 57]. For these most common capillary columns, a high percentage of Methyl end groups results in low polarity columns. In contrast, a higher percentage of cyanopropylphenyl end groups results in high polarity columns. Phenyl end groups lead to strong interactions with the π-system of unsaturated or aromatic compounds. Mixtures of the quoted end groups can be used to generate columns matching to an astonishing amount of separation requirements. These capillary columns are available in different lengths (usually 30 to 60 m), depending on complexity of the samples and speed of analysis. A more detailed overview can be found in recent literature [57]. Other material columns like PLOT (porous layer open tubular), packed, or chiral columns are offering a wide variety of packing materials and accessible polarities, thus increasing the range of selectivities [49, 56, 57]. The programmable oven is used to alter the speed of equilibration, allowing to, e.g., shorten retention times by increasing temperature and thus speeding up repeated equilibration and adding a distillation process to the separation.

Transfer Capillary and Ionization Then, the separated analytes are transferred into the ionization source of the mass spectrometer utilizing a heated transfer capillary to avoid condensation. In most cases, ionization is still carried out by the impact of high-energy electrons on the vaporized molecules in high vacuum (EI, electron ionization) [71, 72]. For more polar compounds, chemical ionization (CI) is used instead. Here, the ionization is accomplished upon reaction of the gaseous molecule with preformed bath gas ions [73]. This method usually leads to higher signal for the intact molecular ion and a lower amount of fragmentation. Other methods like coldEI [74] or methods using ionization techniques at or close to atmospheric pressure are existing, [75, 76] but still are rarely applied in ORA. Mass Analyzer and Generation of Mass Spectra After ionization, the analytes’ ions are separated by the m/z ratio and then detected. Since the number of ions in the sample is proportional to the measured signal over a wide concentration range, the method can be used not only for qualitative but also for quantitative analyses. The standard mass analyzer in GC-MS is a single quadrupole MS offering unit resolution over the full mass range at a high scan speed (typically more than five full spectra per second). Triple-quadrupole mass spectrometers enhance the tandem MS capability, increasing the selectivity. Ion trap instruments can offer higher sensitivities and/or MSn capabilities [72, 77]. Especially when analyzing complex mixtures or many unknowns, high-resolution/high mass precision (HR) mass analyzers offer unique possibilities to assign molecular formulae to the measured m/z values. The isotopic pattern, especially of the molecular ions, contains valuable information enabling to increase confidence in the results. Time-of-Flight (TOF) analyzers [78] and, lately, Orbitrap instruments [79] are emerging, while the “classical” HR GC-MS instruments, sector field mass spectrometers, loose significance in the area of molecule mass spectrometry. Other mass analyzers are rarely coupled to gas chromatographs [80, 81].

37

Organic Residues Analysis (ORA) in Archaeology

1087

While standard GC-MS is used for volatile compounds or compounds volatile after derivatization (usually trimethylsilylation or methylation), compounds with a high content of polymeric compounds are analyzed by PyGC-MS. For instance, good results using GC-MS can be achieved for fatty acids as oil markers [1, 2, 82, 83] or small organic acids as wine or fermentation markers, [1, 51] while resins, suet, and waxes, [37, 84–88] some dyes, [89] sugars and cellulose, [90, 91] proteins, [92, 93] or charred objects [90] require analysis utilizing py-GC-MS to compensate the lower volatility. A more detailed description of detected commodities and applications is found in Sect. 37.9.1.

37.5

Liquid Chromatography (LC) and Liquid Chromatography– Mass Spectrometry (LC-MS)

If compared to its importance in other fields of analytical science, LC-MS is still underrepresented in organic residue analysis in the field of archaeology, although several impressive reports exist [50, 53, 94–101]. Similar to GC, applications of LC with non-mass spectrometer detectors show disadvantages when it comes to analysis of unexpected or very complex mixtures. This is even as modern UV/visible diode array detectors (UV/vis DAD) allow sensitive detection together with (limited) structural evidence utilizing the real-time generation of UV/vis spectra. Less frequently, other detectors like IR (infrared, structural information, but less sensitive), evaporative light scattering detector (ELSD, for non-volatile compounds like polymers), refractive index detector (universal), and others are used. Since a non-destructive detector like UV/vis allows coupling to a second detector, often combinations like LC-UV/vis-MS are realized. An alternative for multiple detectors would be the use of flow splitters behind the column. With liquid chromatography offering an extremely wide range of stationary phases (chromatographic columns, see [102–107]), there is rarely a separation issue which cannot be achieved, with or without derivatization. And, different from gas chromatography, also the mobile phase which now is a liquid can be adjusted over a very broad range. Here, parameters such as pH, polarity, and ions in solution (type and concentration) can be optimized [108]. While the mobile phase composition in GC is usually not changed, it is – together with the choice of the stationary phase – the major parameter which is optimized in LC method development [109–111]. In many applications, gradient chromatography is used. The temperature plays a significant but less prominent role [112, 113]. The numerous applications described in literature depict the good applicability of liquid chromatography for the analysis of a vast variety of compound classes: Like for ORA analysis with GC, usually methods from other fields of research like analytical chemistry and clinical, food, forensics, or pharmaceutical sciences can be adopted to match the needs of archaeologists. Analog to GC, coupling of a mass spectrometer to the liquid chromatograph adds an extremely versatile tool. LC-MS is a very sensitive technology offering access to molecular formulae as well as structural elements in addition to the separation. With

1088

S. Polla and A. Springer

that said, the ionization source is targeted. The most common and most versatile source is the electrospray ionization source (ESI) [114, 115]. Since presence of nonvolatiles and/or surfactants can alter the ionization process, the sample is dissolved in water, methanol, acetonitrile, tetrahydrofuran or similar solvents. Often, low concentrations (usually 0.1–1%) of volatile acids, bases, or buffers (e.g., formic acid, triethylamine, or ammonium acetate, typically depending on LC conditions and/or acidity/basicity of analyte) are added. Often the protonated ([M + H]+) or multiply protonated ([M + nH]n+) ions in positive ion mode or deprotonated ions in negative ion mode can be detected. Separation of ions from counterions is another possibility. The formation of fragment ions is less pronounced, since the analyte molecules are protected inside a small solvent droplet while transferring from atmospheric pressure into the high vacuum of the mass spectrometer. Also, most of these closed-shell ions are more stable than radical species characteristically derived from EI ionization. Therefore, typically tandem mass spectrometers offering MS/MS are used for the analysis of unknowns and for structural evaluation. Molecules with ionic up to moderate polarity of small to very big size can be analyzed, including most of the compound classes relevant for ORA. Among others, peptides, proteins, [97, 116–120] lipids, [118, 121, 122] and small organic molecules [123, 124] can be ionized with high efficiency. Other common methods for LC coupling are atmospheric pressure chemical ionization (APCI), [125, 126] expanding the polarity scale to less polar compounds and atmospheric pressure photo ionization (APPI) [125–127] for the analysis of compounds with an enhanced π-electron system, including aromatic compounds. Other methods are less common, although examples are found in literature [128]. The coupling of the so-called reversed-phase (RP) LC with a comparably non-polar packing material based on SiO2 particles modified with long alkyl chains and a mobile phase system matching to ESI (see above) is most straightforward. In literature, this is the application most frequently referred in ORA. In addition to RP-HPLC, the most promising recent developments on the chromatography side of LC-MS for ORA are as follows: • UHPLC with its dramatically decreased particle sizes (< 2 μm vs. 5–10 μm in standard HPLC) leading to a smaller diffusion between the particles, thus a significant better resolution resulting in a) shorter LC runs or b) higher peak capacity [129–131]. Since methods from standard HPLC can be transferred very easily, the number of applications increase exponentially. • Hydrophilic interaction chromatography (HILIC) with its unique selectivities [132–136]. • Columns with dramatically smaller inner diameter (micro- and nano-LC column), leading to lower flow rates. A variation of the ESI-source, nano-electrospray ionization (NSI), [137] generates better sensitivities. It is typically used for OMICS analyses like the methods relevant for ORA, lipidomics, [138] and proteomics [139]. • The new generation of supercritical fluid chromatography (SFC) [106, 140, 141].

37

Organic Residues Analysis (ORA) in Archaeology

1089

LC-MS, compared to GC-MS, is sometimes advantageous, since usually the derivatization step can be skipped. Especially proteomic analysis allows taxonomy of the lifeform which expressed the protein, sometimes down to the lifeforms’ species. However, a method with good selectivity for free saturated and unsaturated as well as oxidized fatty acids offering high sensitivity in MS detection is still not found until today. The method of LC-MS is used for the analysis of organic dyes, [50, 142, 143] lipids and oils from plants and algae, [144–146] arts, [99] vessels, [147] lamps, [148] or ceramic vessels [100, 149] containing oils or fats and cereals [101]. Also in some cases, the bone collagen used for radiocarbon dating [150] or other proteins/ peptides, for example, from deposits inside of vessels, [97] grave goods, [151] human dental calculus, [152] or paintings, [93, 153–155] are studied using proteomic approaches [44].

37.6

Isotope-Ratio Mass Spectrometry (IRMS)

Isotope-ratio mass spectrometry is a very specialized technique and requires mass spectrometers exclusively designed to determine isotope ratios with very high precision. Often, one isotope has a very high natural abundance, while the others are available in minor or even trace amounts (e.g., 12C: ~ 98.9%, 13C: ~ 1.1%, the radioactive 14C (radiocarbon): traces) [156]. There are three issues to overcome: • First, if the molecular information of organic residues would be maintained, a dramatic loss in intensities would be caused by the different molecular weights and complex isotopic pattern of larger molecules (also complicating the issues caused by overlapping peaks). • Second, some orders of magnitude must be covered for high intensity ion beams down to trace amount ions, leading to multidetector instruments. • Third, changes in the isotopic distributions while excavating the artifact, cleaning, or doing sample preparation procedures must be avoided. Something often hard to overcome is contamination of the artifact before the excavation, leading to the necessity of sample extraction/purification protocols. Also, use of improper or impure solvents in the cleaning procedure might lead to irreversible destruction of artifacts for IRMS. To target the first issue, the molecules are destroyed. Since in ORA only isotopic ratios of elements present in organic molecules are of interest, we can ignore most elements but hydrogen, carbon, nitrogen, oxygen, and sulfur. Therefore, instruments for elemental microanalysis usually are the front-end of standard IRMS. In the elemental analyzer, usually the weighted sample is combusted at high temperatures (> 1400
C for analysis of sulfur) with an excess of oxygen. The resulting gases are ionized and analyzed by a mass spectrometer.

1090

S. Polla and A. Springer

To target the second issue, the mass spectrometers used need to separate all peaks of the isotopic pattern of the respective (oxidized) element, also from isobaric interferences. Therefore, high mass resolution/precision and very precise determination of the ratios between the isotopes of interest are required. This creates high demands for the instrumental setup. Usually the mass spectrometers used are sector field instruments with magnetic sectors utilizing constant magnetic fields and an adjustable multidetector setup. This way, also different detector types like Faraday plates for high intensity ion beams and electron multipliers for lower abundances can be used to overcome detector saturation effects [157]. The use of more detectors also means that a reference sample has been analyzed to compensate, for example, fluctuations in detector response while measuring a sample. This comparison of isotopic ratios to certified standard materials (see, e.g., [158]) then is reported as dX ¼ (Rsample/Rstandard) – 1, with X being the relative element and Rsample and Rstandard being the isotopic ratios of this element for sample and standard [159]. A major issue in IRMS is the limited availability of these standards for most elements, creating the need in production of “new” and homogeneous standards. These must be certified again. There is one notable exception: the standard for nitrogen is atmospheric nitrogen. The third issue, contamination, remains a challenge. One way to overcome this issue is to purify one or more compounds and then analyze it. One example thoroughly discussed in literature is IRMS of the long-term stable collagen found inside a protective shell, bone, not only of interest for radiocarbon dating. There are numerous ways of sample preparation/isolation [150, 159–162] targeting different issues like contamination. A very promising approach is replacement of the elemental analyzer in front of the mass spectrometer by a separation technique like gas chromatography (GC-IRMS or GC-combustion-IRMS (GC-C-IRMS)) or, less often, liquid chromatography (LC-IRMS). This methodology allows, for example, age determination of single compounds purified online. This reduces the risk of contamination. With this procedure, the analysis of compounds which are hardly isolated with standard protocols is possible. GC-IRMS was used, among others, for the analysis of lipids and fats [37, 83, 87, 163–171] from marine and terrestrial life forms. LC-IRMS is less common, with one reason being that for analysis of δ13C organic solvents must be avoided, limiting the number of possible applications [166, 172, 173]. In the following, a short summary of the elements relevant for ORA in archaeology is listed, with a short description of the major applications and citations linking to further information. ORA, except pure radiocarbon dating, requires determination of the isotopic ratios of more than one element to improve validity [27, 164, 174, 175]. Hydrogen: 1H, D (2H), (T, 3H) The parameter determined in ORA is δD (δ2H). No isotopes are more different than 1H and its heavier analogues, deuterium, or the trace isotope tritium. Therefore, almost all isotopic effects in chemistry refer to

37

Organic Residues Analysis (ORA) in Archaeology

1091

hydrogen, and physical constants like the melting point of water vs. D2O differ for almost 4 K. Thus, it is no wonder that rainwater contains significantly less deuterium than sea water, with it having highest concentrations close to the equator. The deuterium content can be used to study dietary differences or climate changes [176–179]. One major issue to overcome is that in the big biomolecules some hydrogens are mobile and can be exchanged rapidly (e.g., carboxylic or amine groups exchange hydrogen with the environment quickly, and even amides or alcohols in rates far below time scales relevant for archaeology). One method is using chambers saturated with water or deuterium oxide to determine the average number of exchangeable protons, thus being able to derive better estimations for δD [177, 180]. Carbon: 12C, 13C, 14C (radiocarbon) For carbon, two ratios are of interest: the δ13C, indicating identity of plants (C3 plants like barley, wheat, or rye; C4 plants typically growing hot and dry regions like millet, sugarcane, or maize); CAM plants like pineapple or some succulents cause differences in δ13C due to differences in the photosynthesis pathway [181]. Also diets (vegan or vegetarian eating mainly C3 plants, C4 plants, CAM plants, partially or full carnivore?) can be monitored [27, 164, 172, 174, 175, 182–184]. δ14C or radiocarbon is used to determine the age of an artifact, a theme at the border of ORA and therefore discussed more in detail elsewhere. [33, 150, 175, 185– 192]Harney, 2019 #291}[164, 193–195]. Nitrogen: 14N, 15N When analyzing plants, the δ15N value allows to draw conclusions about the capability of a plant to use atmospheric nitrogen. Nitrogen-fixing plants like lupines or peanuts show a significantly lower δ15N than non-nitrogen-fixing plants like cereal grasses, taking all nitrogen from the soil. Also, δ15N increases about 0,3% for each step of the food chain (especially notable in marine environment with long food chains), [48] so dietary practices can be analyzed for animals and humans [27, 33, 164, 174, 175, 186, 187]. Oxygen: 16O, (17O), 18O For oxygen, δ18O is the ratio analyzed. It is mainly interesting in fields of research targeting the inorganic isotope ratio analysis of water, where it offers access to paleoclimatic information, when, e.g., carbonate-containing fossils like diatoms are analyzed. Higher concentrations of 18O are found close to the equator, especially the oceans, and less toward the polar regions and more in ocean water than in rain [196]. Differences in δ18O lead to different incorporation of 18O in plant material, which can be followed to target mobility as well as dietary aspects (plants vs. marine or terrestrial animals) aside climate changes. In plant material, sugars and similar compounds like cellulose or lignin are good targets since they have a high oxygen content when compared to the molecular weight, [43] but also proteins like keratin from hair have been studied [176]. For more information, see literature [27, 183, 184, 197]. Sulfur: 32S, (33S), 34S, (36S) For sulfur, usually δ34S is the isotope ratio analyzed, and among the elements noted, it is the least important for ORA. In most applications, the inorganic sulfate is

1092

S. Polla and A. Springer

analyzed, which can be another marker for temperature changes. Nevertheless, there are reports indicating that the analysis of δ34S is significant for dietary changes and mobility [174, 175]. δ34S is the most doubtful case for the elements mentioned. It was shown that it is the only element significantly changing isotope ratios when, e.g., the sample is exposed to weathering, especially UV light [39]. The applications of isotopic analysis in the field of biomolecular archaeology of absorbed organic residues in ceramics are numerous [52, 170, 178, 198–205].

37.6.1 Following Climate, Mobility, and Dietary Changes with Spatial-Resolved IRMS Spatial-resolved IRMS allows to follow changes in climate, mobility, and diets, which are well discussed in literature [27, 43, 197, 206]. The isotopic ratios and the length segment can be transferred to time-resolved information. This can be used for the analysis of climate or dietary changes, mobility, and similar, using mainly δ13C, but also δ15N and δ18O ratios. An example of organic IRMS is a spatial-resolved analysis of mainly δ13C, δ14C, and δ18O in trees. Analysis of the tree rings themselves offers additional confirmation of climate changes and age [43]. Using sub-tree ring (annual) resolution, microtomes or robotic micro-milling allows access to sub-annual climate changes. Although isotope ratios in inorganic carbon are targeted in the next example, it plays a major role in spatial-resolved analysis, and the procedure might be adopted for ORA. Determination of δ18O of carbonates deposited in otholiths using robotic micro-drilling of fishes or along the growth bands of calcareous shell fragments of mollusks or of the apatite contained in teeth gives access to majorly climate changes [43, 183]. Bone was found not to be ideal since diagenesis leads to changes hard to be taken into account [43]. As another alternative, Laser ablation-IRMS was shown to allow high spatial resolution for similar samples with reduced sample preparation for δ14C analysis, [207] a methodology which easily could be used for IR-MS of organic residues.

37.7

Other MS-Based Techniques

(Matrix-assisted) laser desorption-ionization mass spectrometry ((MA)LDI-MS) is somewhat limited to bigger molecules. Nevertheless, recent advances in terms of new matrices and procedures facilitate the direct analysis of fatty acids and other small molecules, [208] and dyes [96, 148, 209–213] among others. But, the number of examples in the field of archaeology is still limited. The technique allows analysis with very limited sample preparation and, due to its high sensitivity, virtually not sample consumption. Also, it was shown that MALDI imaging can prove distribution of dyes in ancient fibers [214] Another important technique using mass spectrometers is secondary ion mass spectrometry (SIMS). Although it has comparably harsh ionization conditions and

37

Organic Residues Analysis (ORA) in Archaeology

1093

thus a high extent of fragmentation for the analyte molecules, there are several examples of successful use [44, 62, 215, 216]. Although (MA)LDI and SIMS only require limited sample preparation when compared to GC- or LC-MS, there is some interest in ionization techniques allowing to limit this step even further or even skip it completely. With the advancement of direct analysis in real time (DART), [217] desorption electrospray ionization (DESI), [218] or DE-MS (direct exposure MS), [89] these direct ionization techniques became available. The number of publications in the field of ORA, nevertheless, is limited [60, 84, 89, 219, 220]. Often the direct ionization technique is followed by a more specific analysis, depending on the results of the “shotgun method.”

37.8

Other Analytical Techniques

37.8.1 Infrared Spectroscopy A method with an only very limited demand in terms of sample preparation is infrared spectroscopy (IR). Especially for visible organic residues, it offers a fast and reliable analysis with little demand on knowledge about the sample. It requires an expert to fully understand the resulting IR spectra, but the presence or absence of characteristic IR bands in some range of wavelengths allows a quick classification of the residue. Compound classes like peptides/proteins, fats/oils/fatty acids, resins/ pitch/tar, and more can easily be distinguished. Then, a more sensitive/informative but usually more discrimination method like GC-MS with a lipid extraction protocol can be matched to the results derived from earlier spectroscopic data. Therefore, many examples of using IR spectroscopy in ORA are found in literature, [42, 44, 62, 84, 90, 147, 206, 216, 221–229] although this rarely is the only method used. In ORA, IR is today considered as a preanalytical approach in order to detect the presence of mineral and organic material in the sample [21]. Raman spectroscopy is utilized in similar applications [46, 60, 90, 230].

37.8.2 aDNA Analysis Sequencing of ancient deoxyribonucleic acids (aDNA) is a powerful tool to analyze, e.g., past population dynamics and migrations patterns [231, 232]. Issues to overcome are the comparably low stability of DNA leading to a large extent of strand breaks in archaeological samples. Since the amount of DNA available can be amplified more than 109-fold using polymerase chain reaction (PCR), it is a very sensitive technique, but especially for ancient DNA with its low amount of intact aDNA molecules, contamination and impurities are an issue which must be overcome. Typically, the aDNA is isolated from teeth or bones after they are cleaned and irradiated with UV to remove and destroy contaminations. An inadvertent amplification of aDNA of soil bacteria or the experimenter still might happen und thus might be misleading [233].

1094

S. Polla and A. Springer

Successful applications of aDNA analysis include DNA analysis to study evolution of maize [234] and its distribution over the Americas [235]. Also, its successful use besides microscopy [236] and proteomics [237] in the study of dental calculus [238] as well as in the emerging field of microbial archaeology was shown [239].

37.8.3 Spot Tests For protein residues, fatty acids, carbohydrates, and other compound classes, spot tests exist. These, together with “standard” tests analyzing phosphate, carbonate, pH, calcium, and iron and other inorganic compounds semi-quantitatively, can be used routinely while doing excavations [2, 82, 240, 241]. The spot tests are low in costs and do not need a special training. Areas of a site with different use and function can be defined, giving access, e.g., to aspects targeting social life and economic activities. Since spot tests can be applied in field, fast decisions which artifact might be interesting for special care and further analysis can be made. More specialized, enzyme-linked immunosorbent assays (ELISA) [242] can help to analyze bigger molecules, mostly proteins [44, 48].

37.8.4 NMR Furthermore, powerful analytical tools like especially nuclear magnetic resonance spectroscopy (NMR) can provide additional information, especially about compound classes in mixtures or definite structural evaluation of purified compounds, although examples in archeological ORA still are few and not reflecting its power [228, 243, 244].

37.9

Linking Science and Humanities: Potential and Limits of the Interpretation of ORA-Based Research on Food Systems and Practices in Archaeology

“Food is an area of human life that is intimately linked with both biology and culture.”[245] The definition of “food as embodied material culture” [246] points up its social and cultural role, making food remains and cooking/storage/transport containers a relevant object of archaeological research. The ubiquity of the biomarkers, which are absorbed by several kinds of archaeological surfaces, materials, and vessels types, [21] combined with the technical methods of extracting and identifying complex mixtures of organic compounds, offer the possibility to obtain new information relating to human activity in the past [15]. This approach allows focusing on a high-resolution site scale and provides at the same time valuable information on macroregional perspectives. Therefore, the identification of the chemical fingerprint of the organic detected substances needs to be applied accordingly to an archaeologically and historically informed context, where it matches

37

Organic Residues Analysis (ORA) in Archaeology

1095

to the compounds and mixtures known to exist in vegetable or animal products likely to have been exploited in the concerned time and geographical space [15, 20]. This method poses some problems, since the compounds are mostly common to a very large range of substances, as well of plant as of animal origin, deriving from different mixed commodities stored and processed in ceramic vessels [247]. The interpretation of the detected compounds must therefore be conducted in close collaboration between scientists and archaeologists and involve possibly also other specialists like palaeobotanists and archaeozoologists, in order to improve the explanatory power of the detected biomarkers. Knowledge of the biochemical composition of commodities and of the alteration processes like heating of fats and oils or burial that can affect and modify them is requested [15, 248, 249]. Degradation, hydrolysis, and oxidation are the most frequent and important transformation processes affecting the conservation of organic material. Degradation and alteration processes strongly affect the survival of polyunsaturated fatty acids in ceramic material [6]. The preserved residues are possibly result, on the one hand, of the mixing of different commodities, on the other hand, even of diagenetic alterations, as they reflect site-specific taphonomic situations, i.e., processes of formation, deposition/discard, and burial to recovery and conservation of the samples (also see Sect. 37.3) [14, 21]. The remaining of residues, in particular, of polar molecules, is also strongly affected by their solubility in water. Alteration of ancient lipids through processing and/or aging makes furthermore the comparison with contemporary food composition unreliable [6]. Garnier [21], by focusing on degradation and taphonomy, also considers the difference between the chemical fingerprint of a modern and an ancient substance. On the other hand, the information on alteration processes according to different heating temperatures can even contribute to the study of cooking practices, distinguishing, for example, between dry heat cooking methods, like frying and grilling, and moist heat cooking methods, like boiling and braising [21]. Laboratory experiments [163, 250–254] and ethnographic approaches [254] have contributed to improve the available database of substances and to get a better understanding of the compositional alteration of organic residues during long-term vessel use and deposition. The relation between preservation of residues and the nature of the specific ceramic absorbing material, i.e., its porosity, has been recently tested [255]. Heron and Evershed [6] also emphasize the limited diagnostic potential of fatty acid profiles. Because of the low information value of n-alkanoic acids from plant and animal origin, [15] the distinction of animal fats from plant oils poses great challenges [256]. Fats and oils differ mostly according to the proportions of a very low range of fatty acids, and they show high variability in composition of fatty acids. The composition of plants and seed oils can be affected by soil, climate, and processing. Animal fats show also differences according to the part of animal from which the fats originate [257]. In order to obtain more specific signatures from the organic mixtures resulting from different food products, [23] GC-MS analyses are often integrated by the stable isotopic approach, in order to identify compound-specific fatty acids

1096

S. Polla and A. Springer

[198]. Compound-specific stable isotopic analysis enables to distinguish ruminant from non-ruminant fats and ruminant carcass fats from dairy fats, due to metabolic differences between the different animals and carbon sources utilized in biosynthesis of different fats types. Isotopic signatures can also to better characterize plant fats, by separating C3 plants (wheat, barley, etc.) from C4 plants (millet) (also see Sect. 37.6). Deal et al. [258] offer a comparative study on GC-MS and isotopic analysis.

37.9.1 Detected Commodities Fatty Acids and Sterols Lipids, mainly detected in the form of free fatty acids and possibly sterols, are the most identified compounds using ORA techniques [6, 15, 20]. A large range of different commodities can be recognized via lipids analysis, including, for example, several plant oils, [94, 199, 259–262] terrestrial animal adipose and milk fats, [163, 165, 201, 247, 263, 264] marine animal fats, [87, 202, 247, 249, 265] resins, [266– 268] birch-bark tar, [12, 269] plant waxes, [204, 247, 252, 270, 271] beeswax, [85, 94, 147, 223, 230, 272–276] palm kernel oil, [259] and palm fruits [277] as well as bitumen [278, 147]. Animal and Dairy Fats Due to the low specificity of the most fatty acids, the presence of animal sterols like cholesterol is a meaningful indicator of terrestrial and marine animal vs. plant fats [94]. The so-called palmitic-to-stearic acid ratio [279] is also used to determine the animal vs. plant origin of undiagnostic fatty acids [165]. Nevertheless, since fatty acids are often degradation products derived from other acids, this quantitative discriminant is not reliable. An integrated approach combining GC-MS with isotopic signature (GC-CIRMS) of the detected compounds allows to distinguish plant oils from animal fats [280] and furthermore between ruminants vs. non-ruminants and to separate between carcass and dairy fats and therefore to identify evidence of ancient dairy products (also see Sect. 37.6) [5, 52, 165, 247]. Plant Lipids and the Detection of Diversity of Vegetable Oils Different classes of plant lipids can be potentially assigned to their sources according to the specific hydrocarbons, ketone, alcohols, and/or fatty acids that they contain [204, 270, 272]. Nevertheless, the strong similarity of the compounds’ profile of the most diffused plant oils poses problems in the detectability and distinction of different consumption and trade goods, by discriminating between various sources of plant oils. Biomarkers’ characteristics for plant oils are saturated (C16:0 and C18:0) and unsaturated fatty acids (e.g., C18:1 and C18:2), hydroxy fatty acids, and dicarboxylic acids (on vegetable oils see also [38]). To seek greater precision in the identification of taxa is therefore a research desideratum [281]. Pecci et al. [260] could prove the presence of castor oil in several Late Antique North African transport containers (spatheia and cylindrical amphorae), attested by means of

37

Organic Residues Analysis (ORA) in Archaeology

1097

ricinoleic acid. Together with (olive) oil, wine, and milk, also specific markers of ben oil (long-chain saturated fatty acids associated with phytosterols), obtained by the Moringa sp., have been identified in four African amphorae of Ostia LIX type found in the Rhône river and could be explained as secondary content [282]. Furthermore, brassica, palm fruit, and palm kernel oil have been identified [277, 280, 283]. Radish and castor oil have been proven to be employed as illuminants in lamps from Egypt [199].

Beeswax and Conifer Exudates (Resin, Pitch, and Tar) Resins and waxes and their use, function (i.e., vessels’ waterproofing), commerce, and economics are also object of study using ORA, especially due to the fact that they are well known and very specific and have a good level of preservation, as nonpolar hydrophobic molecules, like bitumen and waxes, are usually well preserved [94, 147, 205, 267, 273, 275, 284]. Chemical markers of frankincense (Boswellia resin), corresponding to alpha- and beta-boswellic and lupeolic acids and their respective O-acetyled derivatives, have been detected in archaeological samples for Yemen, by GC-MS [285]. Characteristic triterpenoids, i.e., boswellic and tirucallic acids, and their numerous dehydrated and oxygenated derivatives, and novel specific compounds, like Δ2-boswellic acids, together with a series of polyunsaturated and aromatic hydrocarbons have been identified in Southern Belgian incense burners [286]. Wine Wine was, alongside oil and fish sauces, the most important commodity traded and consumed in the ancient world. The clear identification of wine using ORA presents many problems because of bad preservation state of wine and other grape derivative residues, and there is still little consensus on the detection and interpretation of wine markers, like tartaric, gallic, malic, lactic, and syringic acid, in archaeological pottery [1, 51, 281, 287, 288, 289]. Garnier et al. [290] and Pecci et al. [51] stimulated the development of specific protocols for the wine extraction, adopting an experimental approach (see Sect. 37.3.1 for discussion of the protocol). Syringic acid has been used as discriminant for the identification of red vs. white wine [95, 98, 288], and polyphenols indicating red wine can also be detected [261]. Anthocyanin malvidin, a plant pigment giving grapes their red color, has been identified using combined liquid chromatography-tandem mass spectrometry (LC-MS/ MS) [98]. Furthermore, the distinction of different fermentation markers (Fig. 37.3) and their interpretation as microbial degradation/alcoholic beverages, such as palm wine, beer, or fruits, are challenging if specific compounds lack, as sugars disappear very quickly. Recently, cereal-based alcohol production in specialized vessels has been detected by analyzing starches [291]. Fish The detection and identification of lipids derived from ancient marine resources is challenging, because of degradation of polyunsaturated fatty acids [22, 25, 83,

1098

S. Polla and A. Springer

Fig. 37.3 Chromatogram showing typical fermentation markers

247, 292]. Fish/fish sauce has been detected in a Roman Dressel 20 amphora from Arles, according to the presence of abundant animal sterols like cholesterol with cholesterol by-products (cholesten and cholestadien) in association with markers of pinacea resin used for coating ceramic walls [293]. Garnier et al. [4] identified fish and fish sauces through the degradation products of cholesterol, using experimental archaeology and organic analyses.

Cereals Despite of their importance in ancient diets, the evidence of cereals in archaeological contexts is scarce, because of their degradation processes. Cereal lipids have been detected by Hamman and Cramp [294] using an experimental approach, by conducting cooking experiments and analyzing cereal-specific compounds such as alkylresorcinols and plant sterols with highly sensitive methods based on GC-QToF-MS. Miliacin, a biomarker for broomcorn millet, has been identified by GC-MS as specific fatty acid indicating the consumption of millet in North-eastern Medieval Italy [88, 295]. A biomarker approach aiming at the detection of secondary lipid metabolites produced by ergot fungi (genus Claviceps), which are common cereal pests, allowed to identify the original presence of Gramineae and to indirectly establish the use of vessels for cereal storage/processing [101]. Compounds typically associated with plant products could be interpreted as biomarkers of wheat and rye bran in the deposit of a wooden container from Switzerland, while the presence of macrobotanical remains corroborated the results [203] Compound-specific analysis with GC-C-IRMS can show whether the sources of the detected compounds are C3 or C4 plants [204, 296].

37

Organic Residues Analysis (ORA) in Archaeology

1099

37.10 The “Biomarkers Revolution” and the Roman Economy Although ORA has been more frequently applied to the study of the pottery from Prehistoric [297] and North European archaeological contexts, for example, Roman Britain, [298–301] new glimpses on traded goods and consumption patterns in the Mediterranean region in the Roman and Late Roman Period have been provided by the recent spreading of research programs including, or dedicated to, the analysis of the organic residues in transport containers (amphorae) and coarse ware [281, 287]. Nevertheless, concerning the Mediterranean environment, no biomarkers or distribution of biomarkers have yet been identified for several important plant products. Also, no reliable isotopic references exist for animals raised in these regional contexts ([281] on Medieval Sicily).

37.10.1 Amphorae The diffusion of Amphorae as long distance transport containers for wine, olive oil, and fish commodities and for a variety of other wet and dry goods shows a peak from the first century BC to the II century AD, fitting well with the peak of the Mediterranean economic activity [302]. Bevan [302] in his essay on long distance packages from the Classical time to the present, noticed that, “despite considerable regional and chronological variability, there were also important efforts at morphological standardization,” attested even at the end of Antiquity [303]. Is it true also for the carried contents and for the relation between a specific container and its content? Several questions concerning the supplying of the Roman Empire with provincial goods and the economic role played by different regions in the production and export of foodstuff can be studied by using chemical analyses of amphorae’s contents [287]. The ORA-based characterization of amphorae’s contents has therefore a dramatic impact on the study and understanding of the ancient economy, in particular, concerning the mobilization of surplus, the commercial flows, and the trade networks. Did a particular amphora type always correspond to the same content and, if so, do we register changes according to chronological and regional variation? In the last decades, the morphotypological and epigraphic research approaches, in particular, concerning the content attribution to Roman Amphorae, [304] have been enriched by the application of ORA techniques for the identification of the content of the transport containers [18, 19, 51, 281, 287, 290, 305], and a reflection on the value of the application of ORA technique and their economic significance in the identification of amphorae-born commodities have been developed [306]. Notably, the assessment of the use of pitch for coating both wine and oil amphorae has challenged the previous traditional interpretation of visible resins/ pitch layers on amphorae walls as clear indicator of wine as content [261]. Amphorae coatings have been studied by Colombini et al. [283]. Especially re-use/secondary use/recycling are critical issues in the interpretation of ORA-based results [307, 308]. The ORA-based model building in the Roman

1100

S. Polla and A. Springer

economy suffers of these issues. It must consider the eventuality that each container had had its own “biography” of multiple use, being not representative for the whole population of the same type. This is particularly true if we consider that Roman and especially Late Roman containers seem to have been multipurpose [287, 309]. The diversity of traded goods and, in particular, of long-distance traded plant oils has been proved using ORA techniques [260, 282]. Furthermore, the development of aDNA analyses will provide high-resolution results [310, 311]. Ancient DNA gives a very precise insight on the range of products carried by amphorae and on their re-use practices, by including not only the transported goods like oil and wine, but even shedding light on the use of foodstuff preservation methods like herbal additives in Greek Amphorae [312].

37.10.2 Cooking Wares: Eating Habits, Culinary Practices, and Diets The application of ORA techniques for the study of diet, eating habits, culinary practices in terms of function, and use of cooking wares has been largely exploited in the Pre- and Protohistoric archaeology. In particular, cooking ware has been successfully analyzed in order to assess, for example, dairying practices and milk consumption in Neolithic Europe [313] or first dairying in Prehistoric Africa [314]. GC-MS and stable isotopic techniques have been recently integrated by proteomics techniques for the analysis of visible encrustations, allowing a more precise identification of milk and dairy proteins in Neolithic cooking vessels from Çatalhöyük [23]. Rageot et al. [315] use ORA for determining the function of Mediterranean vessels imported to Celtic/Middle European contexts in order to study change in consumption practices in the context of cultural contacts in terms of appropriation and adaptation practices of Greek drinking habits. Kimpe et al. [94] and Romanus et al. [316] analyzed the organic residues of several cooking pots from the Roman and Late Roman period Mediterranean in order to study function and use of the vessels. With the scope of conducting a functional study at the assemblage level, [281] the selection of a significant number of samples (40–60) belonging to one or more morphological groups and/or different chronological phases is undeniable in order to define trends and changes in ceramics function and use [14]. Recently, the study of coarse and cooking dating to the Islamic period using ORA has shown its potential for the understanding of daily on-site practices and urban and rural economy [317] and function of ceramics classes. Pecci et al. [318] open a rare window on food practices in Medieval Cyprus. The analysis of adhering organic residues, like starches granules on ceramics [291] and stone tools [319], can improve specificity in the identification of cooked starchy foods, if combined with chemical analysis of the amorphous residues. Drieu et al. [320] shed new light on post-firing treatment of ceramics with organic animal and vegetable substances, i.e., sheep fats and oak leaves, by adopting a multianalytical and experimental approach on this underestimated aspect, strongly affecting

37

Organic Residues Analysis (ORA) in Archaeology

1101

also the misinterpretation of biomarkers from nonfood residues [6] as information on cooking and consumption practices. This study shows that it is at this stage not possible to distinguish the origin of the molecular signals, i.e., between molecular signals resulting from post-firing treatments and those related to the use of pottery (see also [321] on organic coating of Medieval cooking vessels). Misinterpretations can be avoided by applying ORA analyses in the context of a preliminary in-depth ceramological investigation, using macro- and microscopic observation integrated by organic residue analysis. Garnier et al. [293] also deal with impermeabilization/ coating of amphorae and cooking wares with milk, waxes, bitumen, and resin and the difficulty of distinguishing them from the original content. The use of organic adhesive as glue for repairing pottery [322] as well as for decoration purposes [278] has been attested. Wood used as fuel in low-temperature pit-firing is also a possible source of fatty acids in ceramic vessels [323].

37.11 Perspectives Some of the desiderata suggested by Heron and Evershed [6] remains still valid concerning the need of an analytical protocol for the analysis of absorbed lipids including a large numbers of potsherds; and of integrating multiple lines of evidence (typological, petrological, technological, and interdisciplinary contextual information) with the aim of defining patterns of vessel’s use. The much-needed development of an integrated experimental (ethnoarchaeometrical) and interdisciplinary approach, considering contextual paleoethnobotanical information have been recently pointed out [21, 22, 256, 288, 324]. And vice versa, novel avenues of investigation of past environment and climate can be provided by the isotopic ratio analysis (see Sect. 37.6) of the detected compounds in potsherds [22]). Concerning the development of the applied techniques, beside the results of the proteomics of visible residues, [23] some advances in the use of proteomics on amorphous extracted residues have been done [97]. Dallongeville et al. [44]review the state of the art of proteomics of invisible extracted residues. Another very promising technique is 14C dating. Especially dating of single extracted lipids using GC-MS and/or compound-specific stable carbon isotope determination in combination with the accelerator mass spectrometry (AMS) allows to date the organic residues, i.e., the targeted fatty acids (C16:0 and C18:0) preserved in the samples, [22, 193, 325] by linking the individual compounds to commodities processed within the vessels. This technique has been firstly applied to sedimentary lipids and has been recently employed on lipids absorbed on foodcrusts [326] and bog butters [327]. Within the framework of radiocarbon age determination of fibers, pottery and organic materials, [164, 193, 325, 328, 329] or wine fraud, [330] pottery can be dated according to the extracted fatty acid content. Compound-specific dating has therefore the great and still unexploited potential of improving the chronological resolution of undiagnostic coarse and cooking wares and can provide information of the latest use and function of long-lasting ceramics forms in archaeological contexts.

1102

S. Polla and A. Springer

References 1. Pecci A, Cau Ontiveros MÁ, Garnier N (2013) Identifying wine and oil production: analysis of residues from Roman and Late Antique plastered vats. J Archaeol Sci 40(12):4491–4498. https://doi.org/10.1016/j.jas.2013.06.019 2. Pecci A et al (2013) Understanding residues of oil production: chemical analyses of floors in traditional mills. J Archaeol Sci 40(2):883–893. https://doi.org/10.1016/j.jas.2012.07.021 3. Pecci A, Dominguez-Bella S, Buonincontri MP et al (2018) Combining residue analysis of floors and ceramics for the study of activity areas at the Garum shop at Pompeii. Archaeol Anthropol Sci 10(2):485–502. https://doi.org/10.1007/s12520-017-0573-7 4. Garnier N et al (2018) Looking for ancient fish products through invisible biomolecular residues in the Roman production vats from the Atlantic Coast. J Mari Arch 13:285–328 5. Evershed RP et al (1992) The survival of food residues: new methods of analysis, interpretation and application. In: Pollard AM (ed) New developments in archaeological science. Oxford University Press 6. Heron C, Evershed RP (1993) The analysis of organic residues and the study of pottery use. J Archaeol Method Theory 5:247–284 7. Ribechini E et al (2008) Gas chromatographic and mass spectrometric investigations of organic residues from Roman glass unguentaria. J Chromatogr A 1183(1):158–169. https:// doi.org/10.1016/j.chroma.2007.12.090 8. Fullagar R, Matheson C (2014) Stone tool usewear and residue analysis. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 7062–7065. https://doi. org/10.1007/978-1-4419-0465-2_842 9. Shillito LM et al (2011) Biomolecular and micromorphological analysis of suspected faecal deposits at Neolithic Çatalhöyük, Turkey. J Archaeol Sci 38:1869–1877. https://doi.org/10. 1016/j.jas.2011.03.031 10. Bonaduce I et al (2016) Analytical approaches based on gas chromatography mass spectrometry (GC/MS) to study organic materials in artworks and archaeological objects. Topics Curr Chem 374(1):1–37. https://doi.org/10.1007/s41061-015-0007-x 11. Radini A et al (2017) Beyond food: the multiple pathways for inclusion of materials into ancient dental calculus. Am J Phys Anthropol 162(S63):71–83. https://doi.org/10.1002/ajpa.23147 12. Charters S et al (1993) Quantification and distribution of lipid in archaeological ceramics: implications for sampling potsherds for organic residue analysis and the classification of vessel use. Archaeometry 35(2):211–223. https://doi.org/10.1111/j.1475-4754.1993.tb01036.x 13. Stern B et al (2000) A comparison of methods for es-tablishing fatty acid concentration gradients across potsherds: a case study using LateBronze Age Canaanite amphorae. Archaeometry 42:399–414. https://doi.org/10.1111/j.1475-4754.2000.tb00890.x 14. Dunne J et al (2017) Organic residue analysis and archaeology – guidance for good practice. Historic England, Swindon. 978-1-848-02484-7 (print) 15. Evershed RP (2008) Organic residue analysis in archaeology: the archaeological biomarker revolution. Archaeometry 50(6):895–924. https://doi.org/10.1111/j.1475-4754.2008.00446.x 16. Thornton MD, Morgan ED, Celoria F (1970) The composition of bog butter. Sci Archaeol 2/3: 20–25 17. Condamin J et al (1976) The application of gas chromatography to the tracing of oil in ancient amphorae. Archaeometry 18(2):195–201 18. Formenti F, Hensard A, Tchernia A (1978) Note sur le contenu d’une amphore Lamboglia 2 de l’épave de la Madrague de Giens. Archaeonautica 2:95–110 19. Formenti F, Duthel JM (1996) Analysis of wine and other organics inside amphorae of the Roman period. In: McGovern PE, Fleming SJ, Katz SH (eds) The origins and ancient history of wine. Gordon and Breach, p 79 20. Evershed R (1993) Biomolecular archaeology and lipids. World Archaeol 25(1):74–93 21. Garnier N (2015) Méthodologies d’analyse chimique organique en archéologie. In: de Sousa Morais RML, Morillo Cerdán A, Oliveira C (eds) Archaeoanalytics 2014 – chromatography and DNA analysis in archaeology. Município de Esposende

37

Organic Residues Analysis (ORA) in Archaeology

1103

22. Roffet-Salque M et al (2017) From the inside out: upscaling organic residue analyses of archaeological ceramics. J Archaeol Sci Reports 16:627–640. https://doi.org/10.1016/j. jasrep.2016.04.005 23. Hendy J et al (2018) Ancient proteins from ceramic vessels at Çatalhöyük West reveal the hidden cuisine of early farmers. Nat Commun 9(1):4064. https://doi.org/10.1038/s41467-01806335-6 24. Haslam M et al (2009) Archaeological science under a microscope: studies in residue and ancient DNA analysis in Honour of Thomas H. Loy, vol 30. ANU Press. ISBN: 9781921536847 25. Regert M (2011) Analytical strategies for discriminating archeological fatty substances from animal origin. Mass Spectrom Rev 30(2):177–220. https://doi.org/10.1002/mas.20271 26. Arobba D et al (2014) Palaeobotanical, chemical and physical investigation of the content of an ancient wine amphora from the northern Tyrrhenian Sea in Italy. J Archaeol Sci 45:226–233. https://doi.org/10.1016/j.jas.2014.02.024 27. Chesson LA et al (2014) 14.19 – Stable isotopes in forensics applications. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 285–317. https:// doi.org/10.1016/B978-0-08-095975-7.01224-9 28. Smith GD (2019) Archaeometry and antique analysis | Art and conservation. In: Worsfold P et al (eds) Encyclopedia of analytical science, 3rd edn. Academic, Oxford, pp 99–106. https:// doi.org/10.1016/B978-0-12-409547-2.14377-X 29. Bergeron A (2014) Conservation in archaeological practice. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 1646–1656. https://doi.org/10.1007/978-1-44190465-2_1704 30. Comer DC (2014) Conservation and preservation in archaeology in the twenty-first century. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 1640–1646. https://doi.org/10.1007/978-1-4419-0465-2_532 31. Nardi R (2014) Conservation, restoration and preservation in classical archaeology. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 1666–1674. eISBN 978-1-4419-0465-2. ISBN 978-1-4419-0466-9 32. Jones MD et al (2019) 20,000 years of societal vulnerability and adaptation to climate change in Southwest Asia. WIREs Water 6(2):e1330. https://doi.org/10.1002/wat2.1330 33. Kögel-Knabner I, Amelung W (2014) 12.7 – Dynamics, chemistry, and preservation of organic matter in soils. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 157–215. https://doi.org/10.1016/B978-0-08-095975-7.01012-3 34. Petsch ST (2014) 12.8 – Weathering of organic carbon. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 217–238. https://doi.org/10.1016/ B978-0-08-095975-7.01013-5 35. Evershed RP et al (1995) Formation of long-chain ketones in ancient pottery vessels by pyrolysis of acyl lipids. Tetrahedron Lett 36(48):8875–8878. https://doi.org/10.1016/00404039(95)01844-8 36. Aldeias V et al (2016) How heat alters underlying deposits and implications for archaeological fire features: a controlled experiment. J Archaeol Sci 67:64–79. https://doi.org/10.1016/j.jas. 2016.01.016 37. Nieuwenhuyse OP et al (2015) Tracing pottery use and the emergence of secondary product exploitation through lipid residue analysis at Late Neolithic Tell Sabi Abyad (Syria). J Archaeol Sci 64:54–66. https://doi.org/10.1016/j.jas.2015.10.002 38. Heron C, Evershed RP, Goad LJ (1991) Effects of migration of soil lipids on organic residues associated with buried potsherds. J Archaeol Sci 18(6):641–659. https://doi.org/10.1016/ 0305-4403(91)90027-M 39. Auerswald K et al (2011) Does natural weathering change the stable isotope composition (2H, 13 C, 15N, 18O and 34S) of cattle hair? Rapid Commun Mass Spectrom 25(24):3741–3748. https://doi.org/10.1002/rcm.5284 40. Clementi C et al (2011) Non-invasive and micro-destructive investigation of the Domus Aurea wall painting decorations. Anal Bioanal Chem 401(6):1815. https://doi.org/10.1007/s00216011-5250-6

1104

S. Polla and A. Springer

41. Manfredi M et al (2017) Method for noninvasive analysis of proteins and small molecules from ancient objects. Anal Chem 89(6):3310–3317. https://doi.org/10.1021/acs.analchem. 6b03722 42. Lettieri M, Giannotta MT (2017) Investigations by Ft-Ir spectroscopy on residues in pottery cosmetic vases from archaeological sites in the Mediterranean Basin. Int J Exp Spectroscopic Tech 1:9. https://doi.org/10.35840/2631-505X/8509 43. Patterson WP (2014) 14.11 – Microanalytical isotope chemistry: applications for archaeology. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 145–156. https://doi.org/10.1016/B978-0-08-095975-7.01212-2 44. Dallongeville S et al (2016) Proteins in art, archaeology, and paleontology: from detection to identification. Chem Rev 116(1):2–79. https://doi.org/10.1021/acs.chemrev.5b00037 45. Arevalo R (2014) 15.23 – Laser ablation ICP-MS and laser fluorination GS-MS. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 425–441. https://doi.org/10.1016/B978-0-08-095975-7.01432-7 46. Brambilla L et al (2011) Characterization of fresh and aged natural ingredients used in historical ointments by molecular spectroscopic techniques: IR, Raman and fluorescence. Anal Bioanal Chem 401(6):1827–1837. https://doi.org/10.1007/s00216-011-5168-z 47. Correa-Ascencio M, Evershed RP (2014) High throughput screening of organic residues in archaeological potsherds using direct acidified methanol extraction. Anal Methods 6(5): 1330–1340. https://doi.org/10.1039/C3AY41678J 48. Nigra BT, Faull KF, Barnard H (2015) Analytical chemistry in archaeological research. Anal Chem 87(1):3–18. https://doi.org/10.1021/ac5029616 49. Dettmer-Wilde K, Engewald E (2014) Practical gas chromatography – a comprehensive reference. Springer, Berlin/Heidelberg. Print ISBN 978-3-642-54639-6 Online ISBN 978-3642-54640-2 50. Lech K, Jarosz M (2011) Novel methodology for the extraction and identification of natural dyestuffs in historical textiles by HPLC–UV–Vis–ESI MS. Case study: chasubles from the Wawel Cathedral collection. Anal Bioanal Chem 399(9):3241–3251. https://doi.org/10.1007/ s00216-010-4591-x 51. Pecci A et al (2013) Identifying wine markers in ceramics and plasters using gas chromatography–mass spectrometry. Experimental and archaeological materials. J Archaeol Sci 40(1): 109–115. https://doi.org/10.1016/j.jas.2012.05.001 52. Mottram HR et al (1999) New chromatographic, mass spectrometric and stable isotope approaches to the classification of degraded animal fats preserved in archaeological pottery. J Chromatogr A 833(2):209–221. https://doi.org/10.1016/S0021-9673(98)01041-3 53. Mok HJ et al (2016) A rapid and sensitive profiling of free fatty acids using liquid chromatography electrospray ionization tandem mass spectrometry (LC/ESI-MS/MS) after chemical derivatization. RSC Adv 6(38):32130–32139. https://doi.org/10.1039/C6RA01344A 54. Schlenk H, Gellerman JL (1960) Esterification of fatty acids with diazomethane on a small scale. Anal Chem 32(11):1412–1414. https://doi.org/10.1021/ac60167a011 55. Lovelock JE (1974) The electron capture detector: theory and practice. J Chromatogr A 99:3–12. https://doi.org/10.1016/S0021-9673(00)90840-9 56. Scott RPW (2017) Introduction to analytical gas chromatography, revised and expanded, 2nd edn. CRC Press, Boca Raton. eBook ISBN 9781315274546 ISBN: 978-0-824-70016-4 ISBN 9780367400767 57. Rahman MM et al (2015) Basic overview on gas chromatography columns. In: Analytical separation science. Wiley, pp 823–834. https://doi.org/10.1002/9783527678129.assep024 58. Lísa M et al (2011) Characterization of fatty acid and triacylglycerol composition in animal fats using silver-ion and non-aqueous reversed-phase high-performance liquid chromatography/mass spectrometry and gas chromatography/flame ionization detection. J Chromatogr A 1218(42):7499–7510. https://doi.org/10.1016/j.chroma.2011.07.032 59. Aichholz R, Lorbeer E (1999) Investigation of combwax of honeybees with high-temperature gas chromatography and high-temperature gas chromatography–chemical ionization mass

37

Organic Residues Analysis (ORA) in Archaeology

1105

spectrometry: I. High-temperature gas chromatography. J Chromatogr A 855(2):601–615. https://doi.org/10.1016/S0021-9673(99)00725-6 60. Devièse T et al (2011) First chemical evidence of royal purple as a material used for funeral treatment discovered in a Gallo–Roman burial (Naintré, France, third century AD). Anal Bioanal Chem 401(6):1739. https://doi.org/10.1007/s00216-011-5217-7 61. Reber EA (2014) Gas chromatography-mass spectrometry (GC-MS): applications in archaeology. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 2953–2959. https://doi.org/10.1007/978-1-4419-0465-2_340 62. Kałużna-Czaplińska J et al (2017) The studies of archaeological pottery with the use of selected analytical techniques. Crit Rev Anal Chem 47(6):490–498. https://doi.org/10.1080/ 10408347.2017.1334534 63. Robards K, Haddad PR, Jackson PE (2004) 3 – Gas chromatography. In: Robards K, Haddad PR, Jackson PE (eds) Principles and practice of modern chromatographic methods. Academic, Boston, pp 75–177. https://doi.org/10.1016/B978-0-08-057178-2.50006-6 64. Sobolevsky TG et al (2003) Electron ionization mass spectra and their reproducibility for trialkylsilylated derivatives of organic acids, sugars and alcohols. Eur J Mass Spectrom 9(5): 487–495. https://doi.org/10.1255/ejms.580 65. NIST Standard Reference Database (NIST) (ed) (2017) National Institute of Standards and Technology (NIST): Mass Spectrometry Data Center, National Institute of Standards and Technology, Gaithersburg 66. Wiley: Wiley Registry of Mass Spectral Data (2016) Wiley, Hoboken 67. Stein S (2012) Mass spectral reference libraries: an ever-expanding resource for chemical identification. Anal Chem 84(17):7274–7282. https://doi.org/10.1021/ac301205z 68. Allen F et al (2016) Computational prediction of Electron ionization mass spectra to assist in GC/MS compound identification. Anal Chem 88(15):7689–7697. https://doi.org/10.1021/acs. analchem.6b01622 69. Asgeirsson V, Bauer CA, Grimme S (2017) Quantum chemical calculation of electron ionization mass spectra for general organic and inorganic molecules. Chem Sci 8(7):4879–4895. https:// doi.org/10.1039/C7SC00601B 70. de Vries MS et al (2002) Mass spectrometry of polymers. CRC Press, Boca Raton/London/New York/Washington, DC. ISBN 0-8493-3127-7 71. Vestal ML (2001) Methods of ion generation. Chem Rev 101(2):361–376. https://doi.org/10. 1021/cr990104w 72. Gross JH (2011) Mass spectrometry – a textbook, 2nd edn. Springer, Heidelberg. https://doi. org/10.1007/978-3-642-10711-5 73. Munson MSB, Field FH (1966) Chemical ionization mass spectrometry. I. General introduction. J Am Chem Soc 88(12):2621–2630. https://doi.org/10.1021/ja00964a001 74. Amirav A et al (2008) Gas chromatography-mass spectrometry with supersonic molecular beams. J Mass Spectrom 43(2):141–163. https://doi.org/10.1002/jms.1380 75. Li D-X et al (2015) Gas chromatography coupled to atmospheric pressure ionization mass spectrometry (GC-API-MS): review. Anal Chim Acta 891:43–61. https://doi.org/10.1016/j. aca.2015.08.002 76. van Bavel B et al (2015) Atmospheric-pressure chemical ionization tandem mass spectrometry (APGC/MS/MS) an alternative to high-resolution mass spectrometry (HRGC/HRMS) for the determination of dioxins. Anal Chem 87(17):9047–9053. https://doi.org/10.1021/acs. analchem.5b02264 77. Dass C (2001) Principles and practice of biological mass spectrometry. Wiley, New York 78. Mirsaleh-Kohan N, Robertson WD, Compton RN (2008) Electron ionization time-of-flight mass spectrometry: historical review and current applications. Mass Spectrom Rev 27(3):237–285. https://doi.org/10.1002/mas.20162 79. Mol HGJ, Tienstra M, Zomer P (2016) Evaluation of gas chromatography – electron ionization – full scan high resolution Orbitrap mass spectrometry for pesticide residue analysis. Anal Bioanal Chem 935:161–172. https://doi.org/10.1016/j.aca.2016.06.017

1106

S. Polla and A. Springer

80. Barrow MP, Peru KM, Headley JV (2014) An added dimension: GC atmospheric pressure chemical ionization FTICR MS and the Athabasca Oil Sands. Anal Chem 86(16):8281–8288. https://doi.org/10.1021/ac501710y 81. Laude DA Jr et al (1986) Fourier transform mass spectrometry: recent instrumental developments and applications. Mass Spectrom Rev 5(2):107–166. https://doi.org/10.1002/mas. 1280050202 82. Pecci A, D’Andria F (2014) Oil production in Roman times: residue analysis of the floors of an installation in Lecce (southern Italy). J Archaeol Sci 46:363–371. https://doi.org/10.1016/j.jas. 2014.03.019 83. Oras E et al (2017) The adoption of pottery by north-east European hunter-gatherers: evidence from lipid residue analysis. J Archaeol Sci 78:112–119. https://doi.org/10.1016/j.jas.2016. 11.010 84. Ribechini E et al (2008) An integrated analytical approach for characterizing an organic residue from an archaeological glass bottle recovered in Pompeii (Naples, Italy). Talanta 74(4):555–561. https://doi.org/10.1016/j.talanta.2007.06.026 85. Riedo C, Scalarone D, Chiantore O (2011) Pyrolysis–GC/MS for the identification of macromolecular components in historical recipes. Anal Bioanal Chem 401(6):1761–1769. https:// doi.org/10.1007/s00216-011-5142-9 86. Wei S et al (2011) Analytical characterization of lacquer objects excavated from a Chu tomb in China. J Archaeol Sci 38(10):2667–2674. https://doi.org/10.1016/j.jas.2011.05.026 87. Shoda S et al (2017) Pottery use by early Holocene hunter-gatherers of the Korean peninsula closely linked with the exploitation of marine resources. Q Sci Rev 170:164–173. https://doi. org/10.1016/j.quascirev.2017.06.032 88. Bossard N et al (2013) Distribution of miliacin (olean-18-en-3β-ol methyl ether) and related compounds in broomcorn millet (Panicum miliaceum) and other reputed sources: implications for the use of sedimentary miliacin as a tracer of millet. Org Geochem 63:48–55. https://doi. org/10.1016/j.orggeochem.2013.07.012 89. Andreotti A et al (2004) Characterisation of natural indigo and shellfish purple by mass spectrometric techniques. Rapid Commun Mass Spectrom 18(11):1213–1220. https://doi. org/10.1002/rcm.1464 90. Canevali C et al (2011) A multi-analytical approach for the characterization of powders from the Pompeii archaeological site. Anal Bioanal Chem 401(6):1801–1814. https://doi.org/10. 1007/s00216-011-5216-8 91. Sanjurjo-Sánchez J, Kaal J, Fenollós JLM (2018) Organic matter from bevelled rim bowls of the middle Euphrates: results from molecular characterization using pyrolysis-GC–MS. Microchem J 141:1–6. https://doi.org/10.1016/j.microc.2018.05.001 92. Wei S, Ma Q, Schreiner M (2012) Scientific investigation of the paint and adhesive materials used in the Western Han dynasty polychromy terracotta army, Qingzhou, China. J Archaeol Sci 39(5):1628–1633. https://doi.org/10.1016/j.jas.2012.01.011 93. Orsini S, Parlanti F, Bonaduce I (2017) Analytical pyrolysis of proteins in samples from artistic and archaeological objects. J Anal Appl Pyrol 124:643–657. https://doi.org/10.1016/j.jaap. 2016.12.017 94. Kimpe K, Jacobs PA, Waelkens M (2002) Mass spectrometric methods prove the use of beeswax and ruminant fat in late Roman cooking pots. J Chromatogr A 968(1–2):151–160. https://doi.org/10.1016/S0021-9673(02)00825-7 95. Guasch-Jané MR et al (2004) Liquid chromatography with mass spectrometry in tandem mode applied for the identification of wine markers in residues from ancient Egyptian vessels. Anal Chem 76(6):1672–1677. https://doi.org/10.1021/ac035082z 96. Van Elslande E et al (2008) Analysis of ancient Greco–Roman cosmetic materials using laser desorption ionization and electrospray ionization mass spectrometry. Anal Bioanal Chem 390(7):1873–1879. https://doi.org/10.1007/s00216-008-1924-0 97. Dallongeville S et al (2011) Dealing with the identification of protein species in ancient amphorae. Anal Bioanal Chem 399(9):3053–3063. https://doi.org/10.1007/s00216-010-4218-2

37

Organic Residues Analysis (ORA) in Archaeology

1107

98. Barnard H et al (2011) Chemical evidence for wine production around 4000 BCE in the Late Chalcolithic Near Eastern highlands. J Archaeol Sci 38(5):977–984. https://doi.org/10.1016/j. jas.2010.11.012 99. La Nasa J et al (2015) Novel application of liquid chromatography/mass spectrometry for the characterization of drying oils in art: elucidation on the composition of original paint materials used by Edvard Munch (1863–1944). Anal Chim Acta 896:177–189. https://doi.org/10.1016/j. aca.2015.09.023 100. Blanco-Zubiaguirre L et al (2018) GC–MS and HPLC-ESI-QToF characterization of organic lipid residues from ceramic vessels used by Basque whalers from 16th to 17th centuries. Microchem J 137:190–203. https://doi.org/10.1016/j.microc.2017.10.017 101. Lucejko JJ et al (2018) Long-lasting ergot lipids as new biomarkers for assessing the presence of cereals and cereal products in archaeological vessels. Sci Rep 8(1):3935. https://doi.org/10. 1038/s41598-018-22140-z 102. Žuvela P et al (2019) Column characterization and selection systems in reversed-phase highperformance liquid chromatography. Chem Rev 119(6):3674–3729. https://doi.org/10.1021/ acs.chemrev.8b00246 103. Johnson AR, Vitha MF (2011) Chromatographic selectivity triangles. J Chromatogr A 1218(4):556–586. https://doi.org/10.1016/j.chroma.2010.09.046 104. Zhang Y, Luo H, Carr PW (2012) Silica-based, hyper-crosslinked acid stable stationary phases for high performance liquid chromatography. J Chromatogr A 1228:110–124. https://doi.org/ 10.1016/j.chroma.2011.07.096 105. Arrua RD et al (2012) Review of recent advances in the preparation of organic polymer monoliths for liquid chromatography of large molecules. Anal Chim Acta 738:1–12. https:// doi.org/10.1016/j.aca.2012.05.052 106. Galea C, Mangelings D, Vander Heyden Y (2015) Characterization and classification of stationary phases in HPLC and SFC – a review. Anal Chim Acta 886:1–15. https://doi.org/ 10.1016/j.aca.2015.04.009 107. Chester TL (2012) Recent developments in high-performance liquid chromatography stationary phases. Anal Chem 85(2):579–589. https://doi.org/10.1021/ac303180y 108. Samanidou VF (2015) Basic LC method development and optimization. In: Analytical separation science, pp 25–42. https://doi.org/10.1002/9783527678129.assep002 109. Heinisch S, Rocca JL (2004) Effect of mobile phase composition, pH and buffer type on the retention of ionizable compounds in reversed-phase liquid chromatography: application to method development. J Chromatogr A 1048(2):183–193. https://doi.org/10.1016/j.chroma. 2004.07.022 110. Kalizan R, Wiczling P (2005) Theoretical opportunities and actual limitations of pH gradient HPLC. Anal Bioanal Chem 382(3):718–727. https://doi.org/10.1007/s00216-004-3023-1 111. Jandera P (2006) Can the theory of gradient liquid chromatography be useful in solving practical problems? J Chromatogr A 1126(1–2):195–218. https://doi.org/10.1016/j.chroma. 2006.04.094 112. Guillarme D, Heinisch D, Rocca JL (2004) Effect of temperature in reversed phase liquid chromatography. J Chromatogr A 1052(1–2):39–51. https://doi.org/10.1016/j.chroma.2004. 08.052 113. Wiese S, Teutenberg T, Schmidt TC (2012) A general strategy for performing temperatureprogramming in high performance liquid chromatography – further improvements in the accuracy of retention time predictions of segmented temperature gradients. J Chromatogr A 1222:71–80. https://doi.org/10.1016/j.chroma.2011.12.018 114. Fenn JB et al (1990) Electrospray ionization – principles and practice. Mass Spectrom Rev 9(1):37–70. https://doi.org/10.1002/mas.1280090103 115. Fenn JB (2003) Electrospray wings for molecular elephants (Nobel lecture). Angew Chem Int Ed 42(33):3871–3894. https://doi.org/10.1002/anie.200300605 116. Wysocki VH et al (2005) Mass spectrometry of peptides and proteins. Methods 35(3):211–222. https://doi.org/10.1016/j.ymeth.2004.08.013

1108

S. Polla and A. Springer

117. Cantin GT, Yates JR III (2004) Strategies for shotgun identification of post-translational modifications by mass spectrometry. J Chromatogr A 1053(1–2):7–14. https://doi.org/10. 1016/j.chroma.2004.06.046 118. Zhang X et al (2007) Mass spectrometry-based “omics” technologies in cancer diagnostics. Mass Spectrom Rev 26(3):403–431. https://doi.org/10.1002/mas.20132 119. Aebersold R, Goodlett DR (2001) Mass spectrometry in proteomics. Chem Rev 101(2):269– 296. https://doi.org/10.1021/cr990076h 120. Riley NM, Coon JJ (2016) Phosphoproteomics in the age of rapid and deep proteome profiling. Anal Chem 88(1):74–94. https://doi.org/10.1021/acs.analchem.5b04123 121. Han X, Gross RW (2005) Shotgun lipidomics: electrospray ionization mass spectrometric analysis and quantitation of cellular lipidomes directly from crude extracts of biological samples. Mass Spectrom Rev 24(3):367–412. https://doi.org/10.1002/mas.20023 122. Li X, Franke AA (2011) Improved LCMS method for the determination of fatty acids in red blood cells by LCOrbitrap MS. Anal Chem 83(8):3192–3198. https://doi.org/10.1021/ ac103093w 123. De Vijlder T et al (2018) A tutorial in small molecule identification via electrospray ionizationmass spectrometry: the practical art of structural elucidation. Mass Spectrom Rev 37(5):607–629. https://doi.org/10.1002/mas.21551 124. Regueiro J et al (2013) Development of a LC–ESI-MS/MS approach for the rapid quantification of main wine organic acids in human urine. J Agric Food Chem 61(27):6763–6768. https://doi.org/10.1021/jf401839g 125. Terrier P et al (2011) APCI/APPI for synthetic polymer analysis. Mass Spectrom Rev 30(5): 854–874. https://doi.org/10.1002/mas.20302 126. Cai SS, Syage JA (2006) Comparison of atmospheric pressure photoionization, atmospheric pressure chemical ionization, and electrospray ionization mass spectrometry for analysis of lipids. Anal Chem 78(4):1191–1199. https://doi.org/10.1021/ac0515834 127. Raffaelli A, Saba A (2003) Athmospheric pressure photo ionization mass spectrometry. Mass Spectrom Rev 22(5):318–331. https://doi.org/10.1002/mas.10060 128. Cappiello A et al (2001) New trends in the application of electron ionization to liquid chromatography-mass spectrometry interfacing. Mass Spectrom Rev 20(2):88–104. https:// doi.org/10.1002/mas.1004 129. Mazzeo JR et al (2005) Advancing LC performance with smaller particles and higher pressure. Anal Chem 77(23):460A–467A 130. Grumbach ES, Arsenault JC, McCabe DR (2009) Beginners guide to UPLC – ultraperformance liquid chromatography. Waters Co, Milford. ISBN: 9781879732070 131. Meyer V (2010) The 1000 bar and 24 hour limits of one-dimensional HPLC – graphical representations. Anal Bioanal Chem 396(5):1633–1640. https://doi.org/10.1007/s00216-0093430-4 132. Schwalbe-Herrmann M, Willmann J, Leibfritz D (2010) Separation of phospholipid classes by hydrophilic interaction chromatography detected by electrospray ionization mass spectrometry. J Chromatogr A 1217(32):5179–5183. https://doi.org/10.1016/j.chroma.2010.05.014 133. Jandera P (2011) Stationary and mobile phases in hydrophilic interaction chromatography: a review. Anal Chim Acta 692(1–2):1–25. https://doi.org/10.1016/j.aca.2011.02.047 134. Gika H et al (2012) Hydrophilic interaction ultra performance liquid chromatography retention prediction under gradient elution. Anal Bioanal Chem 404(3):701–709. https://doi.org/10. 1007/s00216-012-6015-6 135. Euerby MR et al (2015) Retention modelling in hydrophilic interaction chromatography. Anal Bioanal Chem 407(30):9135–9152. https://doi.org/10.1007/s00216-015-9079-2 136. Tang D-Q et al (2016) HILIC-MS for metabolomics: an attractive and complementary approach to RPLC-MS. Mass Spectrom Rev 35(5):574–600. https://doi.org/10.1002/mas. 21445 137. Wilm M, Mann M (1996) Analytical properties of the nanoelectrospray ion source. Anal Chem 68(1):1–8. https://doi.org/10.1021/ac9509519

37

Organic Residues Analysis (ORA) in Archaeology

1109

138. Danne-Rasche N, Coman C, Ahrends R (2018) Nano-LC/NSI MS refines Lipidomics by enhancing lipid coverage, measurement sensitivity, and linear dynamic range. Anal Chem 90(13):8093–8101. https://doi.org/10.1021/acs.analchem.8b01275 139. Aebersold R, Mann M (2003) Mass spectrometry-based proteomics. Nature 422(6928):198–207. https://doi.org/10.1038/nature01511 140. Taylor LT (2008) Supercritical fluid chromatography. Anal Chem 80(12):4285–4294. https:// doi.org/10.1021/ac800482d 141. Nováková L et al (2014) Modern analytical supercritical fluid chromatography using columns packed with sub-2 μm particles: a tutorial. Anal Chim Acta 824:18–35. https://doi.org/10. 1016/j.aca.2014.03.034 142. Ribechini E et al (2011) Discovering the composition of ancient cosmetics and remedies: analytical techniques and materials. Anal Bioanal Chem 401(6):1727–1738. https://doi.org/10. 1007/s00216-011-5112-2 143. Degano I et al (2009) Analytical methods for the characterization of organic dyes in artworks and in historical textiles. Appl Spectrosc Rev 44(5):363–410. https://doi.org/10.1080/ 05704920902937876 144. La Nasa J et al (2013) Core shell stationary phases for a novel separation of triglycerides in plant oils by high performance liquid chromatography with electrospray-quadrupole-time of flight mass spectrometer. J Chromatogr A 1308:114–124. https://doi.org/10.1016/j.chroma. 2013.08.015 145. Gritti F, Perdu M-A, Guiochon G (2012) Gradient HPLC of samples extracted from the green microalga Botryococcus braunii using highly efficient columns packed with 2.6μm KinetexC18 core–shell particles. J Chromatogr A 1229:148–155. https://doi.org/10.1016/j.chroma. 2012.01.013 146. Saliu F et al (2011) HPLC–APCI-MS analysis of triacylglycerols (TAGs) in historical pharmaceutical ointments from the eighteenth century. Anal Bioanal Chem 401(6):1785–1800. https://doi.org/10.1007/s00216-011-5179-9 147. Baeten J et al (2010) Application of a multi-analytical toolset to a 16th century ointment: identification as lead plaster mixed with beeswax. Microchem J 95(2):227–234. https://doi. org/10.1016/j.microc.2009.12.005 148. Garnier N et al (2009) Analysis of archaeological triacylglycerols by high resolution nanoESI, FT-ICR MS and IRMPD MS/MS: application to 5th century BC–4th century AD oil lamps from Olbia (Ukraine). Int J Mass Spectrom 284(1):47–56. https://doi.org/10.1016/j.ijms.2009. 03.003 149. Mirabaud S, Rolando C, Regert M (2007) Molecular criteria for discriminating adipose fat and milk from different species by NanoESI MS and MS/MS of their triacylglycerols: application to archaeological remains. Anal Chem 79(16):6182–6192. https://doi.org/10. 1021/ac070594p 150. Cersoy S et al (2019) Toward a versatile protocol for radiocarbon and proteomics analysis of ancient collagen. J Archaeol Sci 101:1–10. https://doi.org/10.1016/j.jas.2018.10.009 151. Yang Y et al (2014) Proteomics evidence for kefir dairy in Early Bronze Age China. J Archaeol Sci 45:178–186. https://doi.org/10.1016/j.jas.2014.02.005 152. Warinner C et al (2014) Direct evidence of milk consumption from ancient human dental calculus. Sci Rep 4:7104. https://doi.org/10.1038/srep07104 153. Dallongeville S et al (2011) Identification of animal glue species in artworks using proteomics: application to a 18th century gilt sample. Anal Chem 83(24):9431–9437. https://doi.org/10. 1021/ac201978j 154. Tokarski C et al (2006) Identification of proteins in renaissance paintings by proteomics. Anal Chem 78(5):1494–1502. https://doi.org/10.1021/ac051181w 155. Vinciguerra R et al (2015) Deglycosylation step to improve the identification of egg proteins in art samples. Anal Chem 87(20):10178–10182. https://doi.org/10.1021/acs.analchem.5b02423 156. Zare RN (2012) Ultrasensitive radiocarbon detection. Nature 482(7385):312–313. https://doi. org/10.1038/482312a

1110

S. Polla and A. Springer

157. Sharp ZD (2014) 15.16 – Stable isotope techniques for gas source mass spectrometry. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 291–307. https:// doi.org/10.1016/B978-0-08-095975-7.01425-X 158. Jochum KP, Enzweiler J (2014) 15.3 – Reference materials in geochemical and environmental research. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 43–70. https://doi.org/10.1016/B978-0-08-095975-7.01403-0 159. Matsubayashi J, Tayasu I (2019) Collagen turnover and isotopic records in cortical bone. J Archaeol Sci 106:37–44. https://doi.org/10.1016/j.jas.2019.03.010 160. Dobberstein RC et al (2009) Archaeological collagen: why worry about collagen diagenesis? Archaeol Anthropol Sci 1(1):31–42. https://doi.org/10.1007/s12520-009-0002-7 161. Maspero F et al (2011) A new procedure for extraction of collagen from modern and archaeological bones for 14C dating. Anal Bioanal Chem 401(6):2019. https://doi.org/10. 1007/s00216-011-5252-4 162. Arslanov KA, Svezhentsev YS (2016) An improved method for radiocarbon dating fossil bones. Radiocarbon 35(3):387–391. https://doi.org/10.1017/S0033822200060392 163. Dudd SN, Evershed RP (1998) Direct demonstration of milk as an element of archaeological economies. Science 282(5393):1478–1481. https://doi.org/10.1126/science.282.5393.1478 164. Craig OE et al (2013) Earliest evidence for the use of pottery. Nature 496:351–354. https://doi. org/10.1038/nature12109 165. Evershed RP et al (2002) Chemistry of archaeological animal fats. Acc Chem Res 35(8):660– 668. https://doi.org/10.1021/ar000200f 166. Dunn PJH, Honch NV, Evershed RP (2011) Comparison of liquid chromatography–isotope ratio mass spectrometry (LC/IRMS) and gas chromatography–combustion–isotope ratio mass spectrometry (GC/C/IRMS) for the determination of collagen amino acid δ13C values for palaeodietary and palaeoecological reconstruction. Rapid Commun Mass Spectrom 25(20): 2995–3011. https://doi.org/10.1002/rcm.5174 167. Papakosta V et al (2015) Extraction and derivatization of absorbed lipid residues from very small and very old samples of ceramic potsherds for molecular analysis by gas chromatography–mass spectrometry (GC–MS) and single compound stable carbon isotope analysis by gas chromatography–combustion–isotope ratio mass spectrometry (GC–C–IRMS). Microchem J 123:196–200. https://doi.org/10.1016/j.microc.2015.06.013 168. Lucquin A et al (2016) Ancient lipids document continuity in the use of early hunter–gatherer pottery through 9,000 years of Japanese prehistory. Proc Natl Acad Sci U S A 113(15):3991– 3996. https://doi.org/10.1073/pnas.1522908113 169. Heron C et al (2015) Cooking fish and drinking milk? Patterns in pottery use in the southeastern Baltic, 3300–2400 cal BC. J Archaeol Sci 63:33–43. https://doi.org/10.1016/j.jas. 2015.08.002 170. Heron C et al (2016) First molecular and isotopic evidence of millet processing in prehistoric pottery vessels. Sci Rep 6:38767. https://doi.org/10.1038/srep38767 171. Papakosta V, Oras E, Isaksson S (2019) Early pottery use across the Baltic – a comparative lipid residue study on Ertebølle and Narva ceramics from coastal hunter-gatherer sites in southern Scandinavia, northern Germany and Estonia. J Archaeol Sci Reports 24:142–151. https://doi.org/10.1016/j.jasrep.2019.01.003 172. Godin J-P, Fay L-B, Hopfgartner G (2007) Liquid chromatography combined with mass spectrometry for 13C isotopic analysis in life science research. Mass Spectrom Rev 26(6): 751–774. https://doi.org/10.1002/mas.20149 173. Raghavan M et al (2010) Amino acid δ13C analysis of hair proteins and bone collagen using liquid chromatography/isotope ratio mass spectrometry: paleodietary implications from intraindividual comparisons. Rapid Commun Mass Spectrom 24(5):541–548. https://doi.org/10. 1002/rcm.4398 174. Hamilton WD et al (2019) ‘Celtic cowboys’ reborn: application of multi-isotopic analysis (δ13C, δ15N, and δ34S) to examine mobility and movement of animals within an Iron Age British society. J Archaeol Sci 101:189–198. https://doi.org/10.1016/j.jas.2018.04.006

37

Organic Residues Analysis (ORA) in Archaeology

1111

175. Etu-Sihvola H et al (2019) The dIANA database – resource for isotopic paleodietary research in the Baltic Sea area. J Archaeol Sci Reports 24:1003–1013. https://doi.org/10.1016/j.jasrep. 2019.03.005 176. Ehleringer JR et al (2008) Hydrogen and oxygen isotope ratios in human hair are related to geography. Proc Natl Acad Sci U S A 105(8):2788–2793. https://doi.org/10.1073/pnas. 0712228105 177. Reynard LM, Hedges REM (2008) Stable hydrogen isotopes of bone collagen in palaeodietary and palaeoenvironmental reconstruction. J Archaeol Sci 35(7):1934–1942. https://doi.org/10. 1016/j.jas.2007.12.004 178. Cramp L, Evershed RP (2014) 14.20 – Reconstructing aquatic resource exploitation in human prehistory using lipid biomarkers and stable isotopes. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 319–339. https://doi.org/10.1016/ B978-0-08-095975-7.01225-0 179. Moffat I (2014) Isotope geochemistry in archaeology. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 4106–4111. https://doi.org/10.1007/978-1-4419-04652_1712 180. Schimmelmann A (1991) Determination of the concentration and stable isotopic composition of nonexchangeable hydrogen in organic matter. Anal Chem 63(21):2456–2459. https://doi. org/10.1021/ac00021a013 181. Cerling TE (2014) 14.12 – Stable isotope evidence for hominin environments in Africa. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 157– 167. https://doi.org/10.1016/B978-0-08-095975-7.01213-4 182. Quade J (2014) 14.10 – The carbon, oxygen, and clumped isotopic composition of soil carbonate in archeology. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 129–143. https://doi.org/10.1016/B978-0-08-095975-7.01211-0 183. Sponheimer M, Cerling TE (2014) 14.21 – Investigating ancient diets using stable isotopes in bioapatites. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 341–355. https://doi.org/10.1016/B978-0-08-095975-7.01222-5 184. Winter-Schuh C et al (2018) Restricted pasturing of domesticated cattle at a late Neolithic settlement in Central Germany. J Archaeol Sci Reports 22:285–297. https://doi.org/10.1016/j. jasrep.2018.07.015 185. Hedges REM, Law IA (1989) The radiocarbon dating of bone. Appl Geochem 4(3):249–253. https://doi.org/10.1016/0883-2927(89)90025-5 186. Piliciauskas G et al (2018) The Corded Ware culture in the Eastern Baltic: new evidence on chronology, diet, beaker, bone and flint tool function. J Archaeol Sci Reports 21:538–552. https://doi.org/10.1016/j.jasrep.2018.08.023 187. Copley MS et al (2003) Direct chemical evidence for widespread dairying in prehistoric Britain. Proc Natl Acad Sci U S A 100(4):1524. https://doi.org/10.1073/pnas.0335955100 188. Jantzi SC, Almirall JR (2014) 14.18 – Trace evidence: glass, paint, soil, and bone. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 271–284. https://doi.org/10.1016/B978-0-08-095975-7.01217-1 189. Hellborg R, Skog G (2008) Accelerator mass spectrometry. Mass Spectrom Rev 27(5):398– 427. https://doi.org/10.1002/mas.20172 190. Jull AJT, Burr GS (2014) 15.20 – Accelerator mass spectrometry. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 375–383. https://doi.org/10. 1016/B978-0-08-095975-7.01429-7 191. Synal H-A (2013) Developments in accelerator mass spectrometry. Int J Mass Spectrom 349–350: 192–202. https://doi.org/10.1016/j.ijms.2013.05.008 192. Higham TFG et al (2018) Radiocarbon dates from the Oxford AMS system: archaeometry Datelist 36. Archaeometry 60(3):628–640. https://doi.org/10.1111/arcm.12372 193. Berstan R et al (2015) Direct dating of pottery from its organic residues: new precision using compound-specific carbon isotopes. Antiquity 82(317):702–713. https://doi.org/10.1017/ S0003598X00097325

1112

S. Polla and A. Springer

194. Hajdas I (2014) 14.3 – Radiocarbon: calibration to absolute time scale. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 37–43. https://doi. org/10.1016/B978-0-08-095975-7.01204-3 195. van Duijn E et al (2014) Automated combustion accelerator mass spectrometry for the analysis of biomedical samples in the low Attomole range. Anal Chem 86(15):7635–7641. https://doi. org/10.1021/ac5015035 196. Bar-Matthews M (2014) 14.9 – History of water in the Middle East and North Africa. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 109–128. https://doi.org/10.1016/B978-0-08-095975-7.01210-9 197. Thompson AH, Wilson AS, Ehleringer JR (2014) 14.23 – Hair as a geochemical recorder: ancient to modern. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 371–393. https://doi.org/10.1016/B978-0-08-095975-7.01227-4 198. Evershed RP et al (1994) Application of isotope ratio monitoring gas chromatography–mass spectrometry to the analysis of organic residues of archaeological origin. Analyst 119(5):909–914. https://doi.org/10.1039/AN9941900909 199. Copley MS et al (2005) Gas chromatographic, mass spectrometric and stable carbon isotopic investigations of organic residues of plant oils and animal fats employed as illuminants in archaeological lamps from Egypt. Analyst 130(6):860–871. https://doi.org/10.1039/B500403A 200. Corr LT, Berstan R, Evershed RP (2007) Optimisation of derivatisation procedures for the determination of δ13C values of amino acids by gas chromatography/combustion/isotope ratio mass spectrometry. Rapid Commun Mass Spectrom 21(23):3759–3771. https://doi.org/10. 1002/rcm.3252 201. Mukherjee AJ et al (2007) Compound-specific stable carbon isotopic detection of pig product processing in British Late Neolithic pottery. Antiquity 81(313):743–754. https://doi.org/10. 1017/S0003598X00095703 202. Craig OE et al (2007) Molecular and isotopic demonstration of the processing of aquatic products in northern European prehistoric pottery. Archaeometry 49(1):135–152. https://doi. org/10.1111/j.1475-4754.2007.00292.x 203. Colonese AC et al (2017) The identification of poultry processing in archaeological ceramic vessels using in-situ isotope references for organic residue analysis. J Archaeol Sci 78:179–192. https://doi.org/10.1016/j.jas.2016.12.006 204. Reber EA et al (2004) Direct detection of maize in pottery residues via compound specific stable carbon isotope analysis. Antiquity 78(301):682–691. https://doi.org/10.1017/ S0003598X00113316 205. Stern B et al (2007) Bulk stable light isotopic ratios in recent and archaeological resins: towards detecting the transport of resins in antiquity? Archaeometry 50(2):351–370. https:// doi.org/10.1111/j.1475-4754.2007.00357.x 206. Schwarcz HP (2014) 14.22 – Human physiology in relation to isotopic studies of ancient and modern humans. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 357–369. https://doi.org/10.1016/B978-0-08-095975-7.01226-2 207. Welte C et al (2016) Laser ablation – accelerator mass spectrometry: an approach for rapid radiocarbon analyses of carbonate archives at high spatial resolution. Anal Chem 88(17): 8570–8576. https://doi.org/10.1021/acs.analchem.6b01659 208. Ling L et al (2018) DBDA as a novel matrix for the analyses of small molecules and quantification of fatty acids by negative ion MALDI-TOF MS. J Am Soc Mass Spectrom 29(4):704–710. https://doi.org/10.1007/s13361-017-1881-y 209. Wyplosz N (2003) Laser desorption mass spectrometric studies of artists’ organic pigments. Faculty of Science (FNWI), Swammerdam Institute for Life Sciences (SILS), Amsterdam 210. Pérez-Arantegui J et al (2009) Colorants and oils in Roman make-ups–an eye witness account. Trend Anal Chem 28(8):1019–1028. https://doi.org/10.1016/j.trac.2009.05.006 211. Ribechini E, Pérez-Arantegui J, Colombini MP (2013) Positive and negative-mode laser desorption/ionization-mass spectrometry (LDI-MS) for the detection of indigoids in archaeological purple. J Mass Spectrom 48(3):384–391. https://doi.org/10.1002/jms.3169

37

Organic Residues Analysis (ORA) in Archaeology

1113

212. Maier MS, Parera SD, Seldes AM (2004) Matrix-assisted laser desorption and electrospray ionization mass spectrometry of carminic acid isolated from cochineal. Int J Mass Spectrom 232(3):225–229. https://doi.org/10.1016/j.ijms.2003.12.008 213. Kirby DP et al (2009) Applications of laser desorption mass spectrometry for the study of synthetic organic pigments in works of art. Int J Mass Spectrom 284(1):115–122. https://doi. org/10.1016/j.ijms.2008.08.011 214. Kramell AE et al (2019) Mapping natural dyes in archeological textiles by imaging mass spectrometry. Sci Rep 9(1):2331. https://doi.org/10.1038/s41598-019-38706-4 215. Oras E et al (2017) MALDI-FT-ICR-MS for archaeological lipid residue analysis. J Mass Spectrom 52(10):689–700. https://doi.org/10.1002/jms.3974 216. Keune K et al (2009) Evaluation of the “added value” of SIMS: a mass spectrometric and spectroscopic study of an unusual Naples yellow oil paint reconstruction. Int J Mass Spectrom 284(1):22–34. https://doi.org/10.1016/j.ijms.2008.10.016 217. Cody RB, Laramée JA, Durst HD (2005) Versatile new ion source for the analysis of materials in open air under ambient conditions. Anal Chem 77(8):2297–2302. https://doi.org/10.1021/ ac050162j 218. Takáts Z et al (2004) Mass spectrometry sampling under ambient conditions with desorption electrospray ionization. Science 306(5695):471–473. https://doi.org/10.1126/science. 1104404 219. Hopkins J, Armitage RA (2012) Characterizing organic residues on ceramics by direct analysis in real time time-of-flight mass spectrometry. In: Lang PL, Armitage RA (eds) Collaborative endeavors in the chemical analysis of art and cultural heritage materials. American Chemical Society, pp 131–142. https://doi.org/10.1021/bk-2012-1103.ch007 220. Modugno F, Ribechini E, Colombini MP (2006) Chemical study of triterpenoid resinous materials in archaeological findings by means of direct exposure electron ionisation mass spectrometry and gas chromatography/mass spectrometry. Rapid Commun Mass Spectrom 20(11):1787–1800. https://doi.org/10.1002/rcm.2507 221. Margaris AV (2014) Fourier transform infrared spectroscopy (FTIR): applications in archaeology. In: Smith C (ed) Encyclopedia of global archaeology. Springer, New York, pp 2890– 2893. https://doi.org/10.1007/978-1-4419-0465-2_343 222. Font J et al (2007) Fourier transform infrared spectroscopy as a suitable technique in the study of the materials used in waterproofing of archaeological amphorae. Anal Chim Acta 598(1): 119–127. https://doi.org/10.1016/j.aca.2007.07.021 223. Ménager M, Azémard C, Vieillescazes C (2014) Study of Egyptian mummification balms by FT-IR spectroscopy and GC–MS. Microchem J 114:32–41. https://doi.org/10.1016/j.microc. 2013.11.018 224. Maravelaki-Kalaitzaki P, Kallithrakas-Kontos N (2003) Pigment and terracotta analyses of Hellenistic figurines in Crete. Anal Chim Acta 497(1):209–225. https://doi.org/10.1016/j.aca. 2003.08.035 225. Izzo FC et al (2013) The study of pitch via gas chromatography–mass spectrometry and Fourier-transformed infrared spectroscopy: the case of the Roman amphoras from Monte Poro, Calabria (Italy). J Archaeol Sci 40(1):595–600. https://doi.org/10.1016/j.jas.2012.06.017 226. Tarquini G, Nunziante Cesaro S, Campanella L (2014) Identification of oil residues in Roman amphorae (Monte Testaccio, Rome): a comparative FTIR spectroscopic study of archeological and artificially aged samples. Talanta 118:195–200. https://doi.org/10.1016/j.talanta.2013. 09.060 227. Lettieri M (2015) Infrared spectroscopic characterization of residues on archaeological pottery through different spectra acquisition modes. Vib Spectrosc 76:48–54. https://doi.org/10.1016/ j.vibspec.2014.12.002 228. Oudemans TFM, Boon JJ, Botto RE (2007) FTIR and solid-state 13C CP/MAS NMR spectroscopy of charred and non-charred solid organic residues preserved in Roman Iron age vessels from the Netherlands. Archaeometry 49(3):571–294. https://doi.org/10.1111/j. 1475-4754.2007.00321.x

1114

S. Polla and A. Springer

229. Shillito LM et al (2009) The use of FT-IR as a screening technique for organic residue analysis of archaeological samples. Spectrochim Acta A Mol Biomol Spectrosc 72(1):120–125. https:// doi.org/10.1016/j.saa.2008.08.016 230. Stacey RJ (2011) The composition of some Roman medicines: evidence for Pliny’s Punic wax? Anal Bioanal Chem 401(6):1749–1759. https://doi.org/10.1007/s00216-011-5160-7 231. Amorim CEG et al (2018) Understanding 6th-century barbarian social organization and migration through paleogenomics. Nat Commun 9(1):3547. https://doi.org/10.1038/s41467018-06024-4 232. Eisenmann S et al (2018) Reconciling material cultures in archaeology with genetic data: the nomenclature of clusters emerging from archaeogenomic analysis. Sci Rep 8(1):13003. https:// doi.org/10.1038/s41598-018-31123-z 233. Pääbo S et al (2004) Genetic analyses from ancient DNA. Annu Rev Genet 38(1):645–679. https://doi.org/10.1146/annurev.genet.37.110801.143214 234. Goloubinoff P, Pääbo S, Wilson AC (1993) Evolution of maize inferred from sequence diversity of an Adh2 gene segment from archaeological specimens. Proc Natl Acad Sci U S A 90(5):1997–2001. https://doi.org/10.1073/pnas.90.5.1997 235. Kistler L et al (2018) Multiproxy evidence highlights a complex evolutionary legacy of maize in South America. Science 362(6420):1309–1313. https://doi.org/10.1126/science.aav0207 236. Radini A et al (2019) Medieval women’s early involvement in manuscript production suggested by lapis lazuli identification in dental calculus. Sci Adv 5(1):eaau7126. https:// doi.org/10.1126/sciadv.aau7126 237. Hendy JEA (2018) Proteomic evidence of dietary sources in ancient dental calculus. Proc R Soc B 285:20180977. https://doi.org/10.1098/rspb.2018.0977 238. Warinner C, Speller C, Collins MJ (2015) A new era in palaeomicrobiology: prospects for ancient dental calculus as a long-term record of the human oral microbiome. Philos Trans Roy Soc B 370(1660):20130376. https://doi.org/10.1098/rstb.2013.0376 239. Warinner C et al (2017) A robust framework for microbial archaeology. Annu Rev Genom Hum G 18(1):321–356. https://doi.org/10.1146/annurev-genom-091416-035526 240. Barba L (2007) Chemical residues in lime-plastered archaeological floors. Geoarchaeology 22(4):439–452. https://doi.org/10.1002/gea.20160 241. Middleton WD et al (2010) The study of archaeological floors: methodological proposal for the analysis of anthropogenic residues by spot tests, ICP-OES, and GC-MS. J Archaeol Method Theory 17(3):183–208. https://doi.org/10.1007/s10816-010-9088-6 242. Aydin S (2015) A short history, principles, and types of ELISA, and our laboratory experience with peptide/protein analyses using ELISA. Peptides 72:4–15. https://doi.org/10.1016/j. peptides.2015.04.012 243. Lambert JB, Shawl CE, Stearns JA (2000) Nuclear magnetic resonance in archaeology. Chem Soc Rev 29(3):175–182. https://doi.org/10.1039/A908378B 244. Spyros A (2016) Liquid-state NMR in cultural heritage and archaeological sciences. In: Webb GA (ed) Modern magnetic resonance. Springer International Publishing, Cham, pp 1–12. https://doi.org/10.1007/978-3-319-28275-6_27-1 245. van der Veen M (2008) Food as embodied material culture: diversity and change in plant food consumption in Roman Britain. J Roman Archaeol 21:83–109. https://doi.org/10.1017/ S1047759400004396 246. Dietler M (2007) Culinary encounters: food, identity, and colonialism. In: Twiss K (ed) The archaeology of food and identity. University of Southern Illinois, pp 218–242 247. Baeten J et al (2013) Molecular evidence for the mixing of meat, fish and vegetables in AngloSaxon coarseware from Hamwic, UK. Archaeometry 55(6):1150–1174. https://doi.org/10. 1111/j.1475-4754.2012.00731.x 248. Evershed RP, Charters S, Quye A (1995) Interpreting lipid residues in archaeological ceramics: preliminary results from laboratory simulations of vessel use and burial. MRS Proc 352:85. https://doi.org/10.1557/PROC-352-85

37

Organic Residues Analysis (ORA) in Archaeology

1115

249. Hansel FA et al (2004) Thermally produced ω-(o-alkylphenyl)alkanoic acids provide evidence for the processing of marine products in archaeological pottery vessels. Tetrahedron Lett 45(14):2999–3002. https://doi.org/10.1016/j.tetlet.2004.01.111 250. Rottländer RCA, Schlichtherle H (1983) Analyse frühgeschichtlicher Gefäßinhalte. Naturwissenschaften 70(1):33–38. https://doi.org/10.1007/bf00365954 251. Charters S et al (1997) Simulation experiments for determining the use of ancient pottery vessels: the behaviour of epicuticular leaf wax during boiling of a leafy vegetable. J Archaeol Sci 24(1):1–7 252. Raven AM et al (1997) Formation of long-chain ketones in archaeological pottery vessels by pyrolysis of acyl lipids. J Anal Appl Pyrol 40–41:267–285. https://doi.org/10.1016/S01652370(97)00036-3 253. Dudd SN, Evershed RP, Gibson AM (1999) Evidence for varying patterns of exploitation of animal products in different prehistoric pottery traditions based on lipids preserved in surface and absorbed residues. J Archaeol Sci 26(12):1473–1482. https://doi.org/10.1006/jasc. 1998.0434 254. Evershed RP (2008) Experimental approaches to the interpretation of absorbed organic residues in archaeological ceramics. World Archaeol 40(1):26–47 255. Drieu L et al (2019) Influence of porosity on lipid preservation in the wall of archaeological pottery. Archaeometry 61(5):1081–1096. https://doi.org/10.1111/arcm.12479 256. McGovern PE, Hall GR (2016) Charting a future course for organic residue analysis in archaeology. J Archaeol Method Theory 23(2):592–622. https://doi.org/10.1007/s10816015-9253-z 257. Mills JS, White R (1987) 3 – Oils and fats. In: Mills JS, White R (eds) The organic chemistry of museum objects. Elsevier, pp 26–40. https://doi.org/10.1016/B978-0-408-11810-1.50008-8 258. Deal M, Morton J, Foulkes E (1991) The role of ceramics among the prehistoric huntergatherers of the Maine-Maritimes region: a view from the New Brunswick interior. In: Deal M (ed) Archaeology of the maritime provinces: past and present research. Council of Maritime Premiers, Maritime Committee on Archaeological Cooperation, New Brunswick, pp 8–40 259. Mazow LB et al (2018) Extraction and analysis of total lipids in late Iron age Bath-shaped basins from the Levant as a means of assessing vessel function. J Archaeol Sci Reports 22: 193–201. https://doi.org/10.1016/j.jasrep.2018.09.015 260. Pecci A et al (2010) Castor Oil at Classe (Ravenna-Italy): residue analysis of some late roman amphorae coming from the port. In: Menchelli S et al (eds) LRCW3. Late Roman coarse wares, cooking wares and amphorae in the Mediterranean archaeology and archaeometry comparison between western and eastern Mediterranean. Archaeopress 261. Romanus K et al (2009) Wine and olive oil permeation in pitched and non-pitched ceramics: relation with results from archaeological amphorae from Sagalassos, Turkey. J Archaeol Sci 36(3):900–909. https://doi.org/10.1016/j.jas.2008.11.024 262. Drieu L, Carver MOH, Craig OO (2018) Commodities carried in amphorae AD 600–1200. New research from Sicily. In: Molinari SGA (ed) I contenitori da trasporto altomedievali e medievali (VIII–XII secolo) nel Mediterraneo. Centri produttori, contenuti, reti di scambio. Archeologia Medievale, pp 269–274 263. Copley MS et al (2005) Processing of milk products in pottery vessels through British prehistory. Antiquity 79(306):895–908. https://doi.org/10.1017/S0003598X00115029 264. Poulain M et al (2016) Dietary practices at the castle of Middelburg, Belgium: organic residue analysis of 16th- to 17th-century ceramics. J Archaeol Sci 67:32–42. https://doi.org/10.1016/j. jas.2016.01.006 265. Copley MS et al (2004) Organic residue evidence for the processing of marine animal products in pottery vessels from the pre-colonial archaeological site of Kasteelberg D east, South Africa: research article. S Afr J Sci 100(5–6):279–283 266. Mills JS, White R (1989) The identity of the resins from the late bronze age shipwreck at Ulu Burun (Kaş). Archaeometry 31(1):37–44. https://doi.org/10.1111/j.1475-4754.1989.tb01054.x

1116

S. Polla and A. Springer

267. Stern B et al (2003) Compositional variations in aged and heated pistacia resin found in late bronze age canaanite amphorae and bowls from Amarna, Egypt. Archaeometry 45(3):457– 469. https://doi.org/10.1111/1475-4754.00121 268. Zareva S, Kuleff I (2010) The application of the derivative IR-spectroscopy and HPLC–ESIMS/MS in the analysis of archaeology resin. Spectrochim Acta A Mol Biomol Spectrosc 76(2):283–286. https://doi.org/10.1016/j.saa.2010.03.021 269. Urem-Kotsou D et al (2002) Birch-bark tar at Neolithic Makriyalos, Greece. Antiquity 76(294):962–967. https://doi.org/10.1017/S0003598X00091766 270. Evershed RP, Heron C, Goad LJ (1991) Epicuticular wax components preserved in potsherds as chemical indicators of leafy vegetables in ancient diets. Antiquity 65(248):540–544. https:// doi.org/10.1017/S0003598X00080145 271. Colombini MP et al (2005) Characterisation of organic residues in pottery vessels of the Roman age from Antinoe (Egypt). Microchem J 79(1):83–90. https://doi.org/10.1016/j. microc.2004.05.004 272. Charters S et al (1995) Evidence for the mixing OF fats and waxes in archaeological ceramics. Archaeometry 37(1):113–127. https://doi.org/10.1111/j.1475-4754.1995.tb00730.x 273. Evershed RP et al (1997) Fuel for thought? Beeswax in lamps and conical cups from Late Minoan Crete. Antiquity 71(274):979–985. https://doi.org/10.1017/S0003598X00085860 274. Regert M et al (2001) Chemical alteration and use of beeswax through time: accelerated ageing tests and analysis of archaeological samples from various environmental contexts. Archaeometry 43(4):549–569. https://doi.org/10.1111/1475-4754.00036 275. Garnier N et al (2002) Characterization of archaeological beeswax by electron ionization and electrospray ionization mass spectrometry. Anal Chem 74(19):4868–4877. https://doi.org/10. 1021/ac025637a 276. Roffet-Salque M et al (2016) Widespread exploitation of the honeybee by early Neolithic farmers. Nature 534(7607):S17–S18. https://doi.org/10.1038/nature18451 277. Copley MS et al (2001) Detection of palm fruit lipids in archaeological pottery from Qasr Ibrim, Egyptian Nubia. Proc R Soc B 268(1467):593–597 278. Connan J et al (2004) Bitumen in early ceramic art: bitumen-painted ceramics from late Neolithic tell Sabi Abyad (Syria)*. Archaeometry 46(1):115–124. https://doi.org/10.1111/j. 1475-4754.2004.00147.x 279. Enser M (1991) Animal carcass fats and fish oils. In: Rossel JB, Pritchard JLR (eds) Analysis of oilseeds, fats and fatty foods. Elsevier Applied Science, London, pp 329–394 280. Steele VJ, Stern B, Stott AW (2010) Olive oil or lard?: distinguishing plant oils from animal fats in the archeological record of the eastern Mediterranean using gas chromatography/ combustion/isotope ratio mass spectrometry. Rapid Commun Mass Spectrom 24(23):3478– 3484. https://doi.org/10.1002/rcm.4790 281. Carver MOH et al (2019) Sicily in transition: new research on early medieval Sicily, 2017– 2018. Fasti On Line Documents & Research, pp 1828–3179 282. Dodinet E, Djaoui D, Garnier N (2015) L’huile de ben identifiée dans quatre amphores africaines de type Ostia LIX provenant d’Arles: difficultés d’interprétation. Ant Afr 51:179–187. https://doi.org/10.3406/antaf.2015.1580 283. Colombini MP, Modugno F, Ribechini E (2005) Organic mass spectrometry in archaeology: evidence for Brassicaceae seed oil in Egyptian ceramic lamps. J Mass Spectrom 40(7):890– 898. https://doi.org/10.1002/jms.865 284. Stern B et al (2008) New investigations into the Uluburun resin cargo. J Archaeol Sci 35(8): 2188–2203. https://doi.org/10.1016/j.jas.2008.02.004 285. Mathe C et al (2004) Characterization of archaeological frankincense by gas chromatography– mass spectrometry. J Chromatogr A 1023(2):277–285. https://doi.org/10.1016/j.chroma.2003. 10.016 286. Baeten J et al (2014) Holy smoke in medieval funerary rites: chemical fingerprints of frankincense in southern Belgian incense burners. PLoS One 9(11):e113142. https://doi.org/ 10.1371/journal.pone.0113142

37

Organic Residues Analysis (ORA) in Archaeology

1117

287. Woodworth M et al (2014) The content of African Keay 25/Africana 3 amphorae: initial results of the CORONAM project. In: ArcheoAnalytics: chromatography and DNA analysis in archaeology. Municipio de Esposende, Esposende. ISBN: 978-989-99468-1-1 288. Garnier N, Valamoti SM (2016) Prehistoric wine-making at Dikili Tash (Northern Greece): integrating residue analysis and archaeobotany. J Archaeol Sci 74:195–206. https://doi.org/10. 1016/j.jas.2016.03.003 289. Drieu L, Rageot M, Wales N, Stern B, Lundy J, Zerrer M et al (2020) Is it possible to identify ancient wine production using biomolecular approaches? Sci Technol Archaeol Res 1–14. https://doi.org/10.1080/20548923.2020.1738728 290. Garnier N et al (2003) Characterization of thermally assisted hydrolysis and methylation products of polyphenols from modern and archaeological vine derivatives using gas chromatography–mass spectrometry. Anal Chim Acta 493(2):137–157. https://doi.org/10.1016/ s0003-2670(03)00869-9 291. Liu L et al (2019) The origins of specialized pottery and diverse alcohol fermentation techniques in Early Neolithic China. Proc Natl Acad Sci U S A 116(26):12767–12774. https://doi.org/10.1073/pnas.1902668116 292. Evershed RP et al (2008) Experimental evidence for the processing of marine animal products in pottery vessels. Archaeometry 50(1):101–113. https://doi.org/10.1111/j.1475-4754.2007. 00368.x 293. Garnier N, Silvino T, Bernal-Casasola D (2011) L’identification du contenu des Amphores. Huile, conserves de poisson et poissage. In: SFECAG. Actes du congrés d’Arles, Marseille, pp 397–416 294. Hammann S, Cramp LJE (2018) Towards the detection of dietary cereal processing through absorbed lipid biomarkers in archaeological pottery. J Archaeol Sci 93:74–81. https://doi.org/ 10.1016/j.jas.2018.02.017 295. Ganzarolli G et al (2018) Direct evidence from lipid residue analysis for the routine consumption of millet in Early Medieval Italy. J Archaeol Sci 96:124–130. https://doi.org/10.1016/j.jas. 2018.06.007 296. Reber EA, Evershed RP (2006) Ancient vegetarians? Absorbed pottery residue analysis of diet in the late woodland and emergent Mississippian periods of the Mississippi Valley. Southeast Archaeol 25(1):110–120 297. Evershed RP et al (1999) Lipids as carriers of anthropogenic signals from prehistory [and discussion]. Philos Trans Roy Soc B 354(1379):19–31 298. Evershed RP et al (2004) Archaeology: formulation of a Roman cosmetic. Nature 432(7013): 35–36. https://doi.org/10.1038/432035a 299. Cramp LJE, Evershed RP, Eckardt H (2011) What was a mortarium used for? Organic residues and cultural change in Iron Age and Roman Britain. Antiquity 85(330):1339–1352. https://doi. org/10.1017/S0003598X00062098 300. Cramp LJE, Evershed RP, Eckardt H (2012) Are you what you grind? A comparison of organic residues from ceramics at two Roman British sites. In: Schruefer-Kolb I (ed) More than just numbers? The role of science in Roman Archaeology. Journal of Roman Archaeology, Portsmouth, pp 93–110. ISBN: 978-1-887829-91-1 301. Cramp L, Evershed R (2015) Reading the residues: chromatographic and mass spectrometric techniques for the reconstruction of artefact use in Roman antiquity. In: Ceramics, cuisine and culture. Oxbow Books, Oxford 302. Bevan A (2014) Mediterranean containerization. Curr Anthropol 55(4):387–418. https://doi. org/10.1086/677034 303. van Alfen PG (2015) The restudy of the LR2 amphoras from the seventh-century Yassiada shipwreck: preliminary evidence for standardization. In: Carlson DN, Kampbell SM, Leidwanger J (eds), Maritime studies in the wake of the Byzantine shipwreck at Yassiada, Turkey, College Station: Texas A&M University Press, pp 17–34 304. Martin-Kilcher S (2011) Fromes d’Amphores et contenu au Haut-Empire, points de repère et questions. In: SFECAG. Actes du congrés d’Arles, Marseille, pp 417–426

1118

S. Polla and A. Springer

305. Garnier N (2007) Analyse de résidus organiques conservés dans des amphores: un état de la question. In: Bonifay M, Tréglia J-C (eds) LRCW 2. Late Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean: Archaeology and archaeometry. BAR Publishing Archaeopress, Oxford, UK. 866pp. ISBN 10:1407300989 ISBN: 9781407300986 306. Bernal D (2014) What contents do we characterise in Roman Amphorae? Methodological and archaeological thoughts on a “trending topic”. In: ArchaeoAnalytics, chromatography and DNA analysis in archaeology. Município de Esposende 307. Peña JT (2007) Roman pottery in the archaeological record. Cambridge University Press, Cambridge. https://doi.org/10.1017/CBO9780511499685 308. Pecci A et al (2017) Use and reuse of amphorae. Wine residues in Dressel 2–4 amphorae from Oplontis Villa B (Torre Annunziata, Italy). J Archaeol Sci Reports 12:515–521. https://doi.org/ 10.1016/j.jasrep.2017.02.025 309. Zanini E (2010) Forma delle anfore e forme del commercio tardoantico: spunti per una riflessione. In: Menchelli SSS, Guiducci MPAG (eds) LRCW3 Late Roman Coarse Wares, cooking wares and amphorae in the Mediterranean: archaeology and archaeometry. Comparison between western and eastern Mediterranean, vol I. BAR International 310. Hansson MC, Foley BP (2008) Ancient DNA fragments inside classical Greek amphoras reveal cargo of 2400-year-old shipwreck. J Archaeol Sci 35(5):1169–1176. https://doi.org/10. 1016/j.jas.2007.08.009 311. Milanesi C et al (2011) Microscope observations and DNA analysis of wine residues from Roman amphorae found in Ukraine and from bottles of recent Tuscan wines. J Archaeol Sci 38(12):3675–3680. https://doi.org/10.1016/j.jas.2011.08.038 312. Foley BP et al (2012) Aspects of ancient Greek trade re-evaluated with amphora DNA evidence. J Archaeol Sci 39(2):389–398. https://doi.org/10.1016/j.jas.2011.09.025 313. Cramp LJE et al (2014) Neolithic dairy farming at the extreme of agriculture in northern Europe. Proc R Soc B 281(1791):20140819. https://doi.org/10.1098/rspb.2014.0819 314. Dunne J et al (2012) First dairying in green Saharan Africa in the fifth millennium BC. Nature 486:390–394. https://doi.org/10.1038/nature11186 315. Rageot M et al (2019) New insights into Early Celtic consumption practices: organic residue analyses of local and imported pottery from Vix-Mont Lassois. PLoS One 14(6):e0218001. https://doi.org/10.1371/journal.pone.0218001 316. Romanus K et al (2007) An evaluation of analytical and interpretative methodologies for the extraction and identification of lipids associated with pottery sherds from the site of Sagalassos, Turkey. Archaeometry 49(4):729–747. https://doi.org/10.1111/j.1475-4754.2007. 00332.x 317. Walker BJ et al (2017) Residue analysis as evidence of activity areas and phased abandonment in a Medieval Jordanian Village. J Islam Archeol 4(2):217–248. https://doi.org/10.1558/jia. 35274 318. Pecci A et al (2015) Preliminary results of the organic residue analysis of 13th century cooking wares from a household in Frankish Paphos (Cyprus). Sci Technol Archaeol Res 1(2):99–105. https://doi.org/10.1080/20548923.2016.1183960 319. Venditti F et al (2019) Animal residues found on tiny Lower Paleolithic tools reveal their use in butchery. Sci Rep 9(1):13031. https://doi.org/10.1038/s41598-019-49650-8 320. Drieu L, Lepère C, Regert M (2019) The Missing Step of Pottery chaîne opératoire: considering post-firing treatments on ceramic vessels using macro- and microscopic observation and molecular analysis. J Archaeol Method Theory 1–25. https://doi.org/10.1007/s10816-01909428-8 321. Salvini L, Pecci A, Giorgi G (2008) Cooking activities during the Middle Ages: organic residues in ceramic vessels from the Sant’Antimo Church (Piombino-Central Italy). J Mass Spectrom 43(1):108–115. https://doi.org/10.1002/jms.1283 322. Charters S et al (1993) Identification of an adhesive used to repair a Roman Jar. Archaeometry 35(1):91–101. https://doi.org/10.1111/j.1475-4754.1993.tb01025.x

37

Organic Residues Analysis (ORA) in Archaeology

1119

323. Reber EA et al (2019) Lipid residues from low-fired pottery. Archaeometry 61(1):131–144. https://doi.org/10.1111/arcm.12403 324. Dunne J et al (2019) Reconciling organic residue analysis, faunal, archaeobotanical and historical records: diet and the medieval peasant at West Cotton, Raunds, Northamptonshire. J Archaeol Sci 107(58–70):0305–4403 325. Stott AW et al (2003) Direct dating of archaeological pottery by compound-specific 14C analysis of preserved lipids. Anal Chem 75(19):5037–5045. https://doi.org/10.1021/ ac020743y 326. Heron C, Craig OE (2015) Aquatic resources in Foodcrusts: identification and implication. Radiocarbon 57(4):707–719 327. Smyth J et al (2019) Four millennia of dairy surplus and deposition revealed through compound-specific stable isotope analysis and radiocarbon dating of Irish bog butters. Sci Rep 9(1):4559. https://doi.org/10.1038/s41598-019-40975-y 328. Jull AJT et al (2004) New frontiers in dating of geological, paleoclimatic and anthropological applications using accelerator mass spectrometric measurements of 14C and 10Be in diverse samples. Glob Planet Chang 41(3):309–323. https://doi.org/10.1016/j.gloplacha.2004.01.014 329. Jull AJT, Burr GS (2014) 14.4 – Radiocarbon: archaeological applications. In: Holland HD, Turekian KK (eds) Treatise on geochemistry, 2nd edn. Elsevier, Oxford, pp 45–53. https://doi. org/10.1016/B978-0-08-095975-7.01205-5 330. Fahrni SM, Fuller BT, Southon JR (2015) Angel’s share combats wine fraud: 14C dating of wine without opening the bottle. Anal Chem 87(17):8646–8650. https://doi.org/10.1021/acs. analchem.5b01998

Towards Preventive Conservation of Stone Artefacts in Historical Gardens by Decay Monitoring

38

Cristiano Riminesi, Rachele Manganelli Del Fá, Silvia Vettori, Fabio Tarani, and Piero Tiano

Contents 38.1 38.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.2.1 The Villa Guicciardini Historical Garden . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.2.2 Diagnostic Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1122 1124 1124 1128 1131 1133 1135

Abstract

While scheduled maintenance is a common practice for the institutions devoted to safeguarding and conserving cultural heritage assets, the prediction of the decay process on stone surfaces is a new challenge in the field. Monitoring is the first, essential step to empower predictive conservation, and technical solutions involving sensors and measurement systems are being widely proposed and tested with the ultimate aim of relating the measured parameters to the decay phenomena. During the activity, we worked on defining a maintenance protocol for stone statues located in historical gardens based on prediction. The final goal is to reduce the degradation risk of cultural heritage assets by means of a smarter planning of conservation works and to optimize budget by avoiding expensive emergency restorations. The paper reports the results of a 5-year-long activity on two marble statues exposed in a historical garden in the surroundings of Florence,

C. Riminesi (*) · R. M. Del Fá · S. Vettori · P. Tiano Institute of Heritage Science – CNR, National Research Council, Florence, Italy e-mail: [email protected]; [email protected]; [email protected]; [email protected] F. Tarani National Research Council, Institute of Heritage Science, Sesto Fiorentino (Firenze), Italy e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_38

1121

1122

C. Riminesi et al.

focusing on the criteria for defining a preventive conservation protocol. The statues are representative of typical exposure conditions in a sub-urban area, and the assessed alteration/decay phenomena are common to many similar artefacts located in historical gardens. Keywords

Outdoor stone artefacts · Diagnostic, conservatin and monitoring · Preventive conservation · ICT for cultural heritage

38.1

Introduction

It is nowadays ascertained that the conservation of heritage buildings and artworks requires preventive maintenance strategies so as to increase the effectiveness and durability of restoration works and to improve their quality and economic sustainability. The same paradigm can be applied to the conservation of historical gardens, suggesting in this case a comprehensive approach including cultural heritage assets and vegetation. Such top-level goal of comprehensively safeguarding the artistic, architectural, and environmental structures is very ambitious, since it involves heterogeneous disciplines and interdisciplinary team work [1]. Public awareness has been growing in the last years on the value of cultural heritage assets and on the importance of their preservation. For monuments and historical buildings, as for human health, prevention is better than cure: conservators know well that systematic prevention on cultural heritage assets is to be preferred over emergency treatments on individual items after damage has occurred. However, the former requires suitable means and methods, because, whereas the causes of damage to heritage are well understood, methods to systematically address them are still fuzzily defined and subject of research. It is in this scope that ICCROM (International Centre for the Study of the Preservation and Restoration of Cultural Property) continues to develop and share tools that respond to the hows and whys of preventive conservation on buildings. These tools, including methods for risk management and reorganization of the maintenance policy, are being implemented in collaboration with leading institutions in the field. For what specifically concerns historical buildings, new technical solutions are continuously provided, acknowledging that the conservation processes strongly benefit from a proper characterization of materials, modes of construction and from a thorough understanding of the decay phenomena. Such knowledge shall be gained through a number of analyses performed during diagnostic campaigns, where non-destructive tests are to be preferred. In addition, computational tools may also be integrated in the workflow to better assess the stress state of the building and to identify potential weaknesses. When tailoring predictive maintenance on historical buildings and artefacts, some additional caveats must be faced. The current trend in research is to provide the most suitable maintenance policy by combining traditional methods with innovative

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1123

technologies: the goal is to get a thorough knowledge of the assets and of the decay modes and to exploit it so as to achieve a desired state of conservation. Each restoration project must hence be understood as an interdisciplinary exchange, involving different specialists who work in synergy to achieve an overall deeper understanding of the conservation issues, while identifying and evaluating the suitability and durability of the proposed solutions [2–7]. As a matter of fact, it is impossible to define a universally ideal intervention, and each specific case must be tackled both by preliminary studies through laboratory tests – thus evaluating efficacy and durability – and by on-field diagnostic activities performed on pilot areas. This approach outlines in advance advantages and drawbacks of the chosen restoration techniques, assuring that the intervention is tested on the real material and state of conservation of the target [8]. However, it might be challenging to locate a pilot area sufficiently large and representative of the whole artifact. Monitoring these pilot areas over time through non-destructive techniques allows to verify durability and effectiveness of the proposed intervention procedure. The importance of continuous monitoring lies in that all works of art, including monuments, have a “safe” allowed response time from the beginning of decay during which it is still possible to recover the original conditions. The threshold before the damage becomes permanent is difficult to determine, but it can be estimated by models based on experimental parameters strictly related to damage risk [9]. Evidently, the intervention threshold must be set well below the damage threshold, since the effects of reversible decay which nevertheless is allowed to progress are more complex to be neutralized [10, 11]. Concerning stone artifacts, the first step is to assess the current state of conservation through a set of parameters and to record their drift over time by means a monitoring system/method based on non-destructive and portable techniques. The assessment should be performed on the surface/structure both in the initial conditions and after each applied conservation treatment, be it cleaning, protecting, or strengthening. In order to systematize such procedure, the most suitable parameters must be selected and the corresponding quantitative diagnostic techniques defined. For stone artefacts, values to be controlled and monitored over time include surface water absorption, color variation, mechanical properties, and the comparison of macro-photos. Besides being strongly related to the intrinsic properties of stone, they also prove to be a useful tool in two different applications: on the one hand, they allow a quantitative evaluation of the performance of applied products in terms of efficacy and durability; on the other hand, their rate of variation, determined through monitoring campaigns, is a fundamental datum to elaborate a maintenance schedule based on predictive models. In this paper, we describe our activity in Villa Guicciardini Corsi Salviati, which aimed at defining the biodegradation thresholds for the surfaces of two marble statues. In order to do this, we studied and developed techniques and monitoring methods based on portable and non-destructive instrumentation and low-cost WiFi sensors. The results are addressed to the owners of the Villa and, more generally, to all institutions, be they public or private, in charge of the care and preservation of cultural heritage assets. The final goal is to provide these with

1124

C. Riminesi et al.

insights on the damage thresholds, so as to empower an efficient maintenance schedule

38.2

Materials and Methods

The garden of Villa Guicciardini Corsi Salviati offers an emblematic instance of Italian Renaissance garden, and its characteristics lead to conservation challenges and suitable monitoring and conservation techniques that used may be easily extended or adapted to a wide ensemble of similar gardens and are described hereafter in Sects. 38.2.1 and 38.2.2.

38.2.1 The Villa Guicciardini Historical Garden The garden of Villa Guicciardini Corsi Salviati is a classical example of Renaissance garden (Fig. 38.1), decorated with statues and furniture crafted out of a variety of stone types such as marble, sandstone and limestone, and terracotta. Each of them is affected by different mechanism of alteration/decay, affected by both its chemical composition and its exposure to atmospheric phenomena (Fig. 38.2). Ironically, previous restorations have shown, due to negligence or scarce knowledge, to worsen the resistance to decay, an example thereof being displayed in Fig. 38.3: the metal

Fig. 38.1 Aerial view of the historical garden and position of the selected statues

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1125

Fig. 38.2 Details of metal joints added during previous restorations

joints added during previous restorations have eventually led to an increased decay in the joints areas, probably due to concentration of mechanical stress and chemical degradation. The environment of the garden favors the biodegradation of surfaces of inorganic materials such as marble and sandstone, depending on their mineral composition and surface morphology (especially roughness and porosity). The predominant “opportunistic” microbial communities will be those capable of selectively growing on materials, while the overall biodiversity will be increased by favorable environmental conditions and by the abundance and variety of minerals offered by the artefact (Fig. 38.4). Finally, further causes of decay and damage are accidents and vandalism, far from unusual on artefacts exposed in public areas. A deeper understanding of the relationships between alteration phenomena and natural and anthropic agents is enabled by a continuous monitoring over time of the material chemical-physical and biological parameters and of environmental conditions, including, e.g., color, water absorption, and vitality of the biological patina. Two different patterns are used in the monitoring of such values: some of them are manually sampled a finite number of times (spot measurements) using manually operated, portable non-destructive devices; others are continuously logged by sensors, e.g., temperature, air humidity ratio, rainfall, wind speed, and solar radiation. For what concerns the four statues surrounding the central fountain, they show superficial degradation and an overall structural damage due to the iron and

1126

C. Riminesi et al.

Fig. 38.3 Examples of biodegradation promoted by green environment

Fig. 38.4 Views of the Doryphoros: initial state (a), after cleaning (b) and final (c)

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1127

Fig. 38.5 Views of the Warrior: initial state (a), after cleaning (b) and final (c)

Fig. 38.6 Deployment of the wireless sensors (center) and, clockwise, details of weather station, surface temperature sensor, pyranometer and rain gauge

copper elements added in time to join the original artefact to the new integrations. Two of them have been chosen as case study: the Doryphoros (Fig. 38.5) and the Warrior (Fig. 38.6); they face each other across the fountain, and their line-ofsight is in the east-west direction. Their surface is rough and weak due to

1128

C. Riminesi et al.

intracrystalline ligand loss, and the marble pieces features cracks near the joints [1]. The restoration policy followed an easy-intervention principle; each statue has been cleaned on half of the surface both on the front and rear side (Figs. 38.4b and 38.5b). After the cleaning, four test areas, one on each side of each statue, have been defined and treated with a brush-spread protective, which features desired traspirability, reversibility, and stability. The product is based on organosiloxane oligomers with a low specific weight and thus a high penetrating power.

38.2.2 Diagnostic Tools Non-destructive techniques (NDT) comprise a wide group of analytic methods used to estimate the properties of a material without affecting its integrity and are thus particularly suited to cultural heritage assets due to their uniqueness and visibility. The most common NDTs used on outdoor artworks include ground-penetrating radar, ultrasonic pulse velocity, remote visual inspection, imaging techniques (VIS, UV, UVr, IR, VIL, IRFC, UVFC), infrared reflectography, light-scattering techniques, photographic techniques for 3D modeling, eddy-current testing (for metal artworks), evanescent field dielectrometry, sponge test method, and colorimetric measures. Some of these techniques are used in the reported activity, together with several sensors to record environmental parameters close to the statues: air temperature and humidity ratio, temperature on the marble surface, wind speed, solar radiation, and rainfall. These sensors belong to a wireless network linked to an IT platform which stores the collected data and allows the management of sensor parameters such as sampling time and download frequency.

Water Absorption Test by the Contact Sponge Method The contact sponge method is a technique to measure the water absorption rate of porous surfaces, and it is used to evaluate the performance of water-repellent treatments. The method, as defined by the Italian technical norm UNI 11432:2011 [12], consists in applying a soaked sponge with a controlled pressure on the surface to be tested [13, 14]. The sponge is soaked (dripping must be avoided) and its weight recorded; then it is applied and kept in contact for 1 to 2 min – depending on the porosity of the receiver – and finally weighed again: the weight difference is a measure of the amount of absorbed water. Adenosine Triphosphate (ATP) Measurements The bio-luminometer is a device which allows to quickly evaluate the microbial contamination of a surface or liquid by means of a bioluminescent reaction. A sampling probe is rubbed on the tested surface and collects, if present, biological material. The reactants contained in the probe release the ATP, an intracellular molecule of all living organisms. The quantity of ATP released is then evaluated by measuring the intensity of bioluminescence produced during the reaction with the

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1129

luciferase enzyme (firefly luciferase reaction [15]: ATP + Luciferin + O2 ! Oxyluciferin + AMP + CO2 + pyrophosphate + light). The amount of light produced, expressed in relative light units (RLU), is directly proportional to the number of living organisms present in the sample: the higher the microbial contamination, the higher the light produced and the value measured by luminometer. In the scope of cultural heritage, this simple and portable device is very useful to detect microbial growth on many types of surfaces, and it is used both to check the biocide/disinfectant efficiency over time and to define, during biocide dose planning, the minimal inhibitory concentration needed to control biological deterioration. In the show application, the measurements have been performed by the portable bio-luminometer 3M Clean-Trace™ NG completed with accessories.

Colorimetric Measurements Quantitative definition of the color of an object is of utmost importance in the scope of cultural heritage conservation, since color is one the most evident characteristics of an artefact, and its variation must be controlled when the item is cleaned, restored, or treated. A colorimeter, after a proper calibration, processes the reflectance spectrum of the tested surface and returns its representative chroma coordinates (L*a*b), which offer an easier framework to evaluate color drift with respect to the more common HSL (hue-saturation-luminosity) coordinates. Namely, measurements take place in the CIELAB color space [16], which is represented on three axes: L*, a*, and b*. In particular, the L* coordinate relates to the luminosity of the color ranging from 0 (black) to 100 (white), whereas variations on the a* and b* axes represent changes in redness-greenness (high a* is redder and low a* is greener) and in yellowness-blueness (high b* is more yellow and low b* is more blue); both axes form a plane orthogonal to L* and do not have specific numerical limits [17]. In the reported activity, the measurements have been carried out with a portable colorimeter (Konica Minolta) with a CM-700d measuring head (8-mm-diameter viewing area), under the following conditions: diffuse illumination geometry with the specular component included in 0 (d/0 ) relative to normal, illuminant D65, and observer 10 . Three readings were taken at different randomly selected zones on each target area. Wireless Sensor Network A wireless sensor network has been deployed so as to monitor a set of environmental parameters. In particular, the network (some details are shown in Fig. 38.6) consists of the following: • A weather station which logs air temperature and humidity ratio, rainfall, and wind speed • Contact sensors which log surface temperature on the statues • A pyranometer which logs solar irradiance in the UV-A spectrum (314–500 nm)

1130

C. Riminesi et al.

Temperature and Air Humidity The weather station features a capacitive humidity sensor with platinum electrodes and a high stability hygroscopic dielectric and a PT100 1/3 DIN resistance thermometer for measuring temperature, completed by a calibration kit. The temperature contact sensors placed on the statues are devices designed with extremely small dimensions, with diameters ranging between 0.46 mm and 3.20 mm depending on the specific application. The thermistor (NTC) offers high accuracy and sensitivity (both  0.1  C) and excellent mechanical resistance. The small dimensions ensure a low thermal capacity that does not alter the measurement and allow a response time shorter than 1 s. These sensors are very suited to measure surface temperature on artefacts when estimating condensation risk. Rain Gauge The rain gauge is contained in a cylinder made of UV-resistant plastic. The rain is collected on a surface of 100 cm2 and is conveyed into a calibrated tilting basin which toggles its position when 0.2 mm of collected rain are reached, thus closing an electrical contact that is acquired by a data logger for subsequent counting. The plastic container has been designed to minimize the impact of wind on the inside, so as to avoid false toppling of the tilting basin. Anemometer Wind speed is measured by a two axes ultrasonic anemometer, getting both wind direction and speed. Such device has some advantages in comparison to traditional mechanical ones: it does not require calibration, and having no moving parts, it reduces maintenance costs. The meaningfulness of measuring intensity and direction of airflow lies in the possible correlation with the deposition of pollutants, which in turn affects stone surfaces. The device can easily be connected to data acquisition systems and integrated in the sensors network, since both analogue and digital outputs are available. UV-A Sensor One of the main causes of performance degradation for protective products is exposure to light and radiation, the most damaging wavelengths lying in the ultraviolet (UV) part of the spectrum. Therefore, a measurement of the incident radiation light in the VIS and UV ranges can provide useful indications to extend product performance. With this aim, we installed a sensor which provides two analogue voltage signals proportional to visible and UV-A irradiances or, in alternative, the μW/lumen ratio, considered an important indicator in the UNI 10829 standard [17] (the usefulness of such ratio lies in that it is independent from the distance to the light source). Internet Connection The system is solar powered and can be queried from the web, thanks to a reserved cloud interface. The web application allows to access to the data stored in the monitoring system, to display it in tabular or graphic form, to view real-time

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1131

measurements, and to configure operating parameters of each data logger such as the sampling time interval.

38.3

Results

The activity has been carried out between February 2015 and April 2017. Firstly, we have assessed the current state of conservation of the statues by micro-invasive analysis [1–3] by mineralogical and petrographic tests on stone samples and biological investigation. Secondly, the statues were cleaned on half of their surface on both sides by mechanical methods (brush and scratcher), and an area spanning 20  15 cm was treated on each of the four cleaned surfaces (one on each side – face and rear – of each statue) with the protective product. Overall, four sets of pilot areas were selected, one on each side (front-rear) and half (left-right) of each statue. Each set consists of three areas, all measuring 20  15 cm: • As is, not-cleaned and not-treated (XX) • Cleaned and not-treated (CX) • cleaned and treated (CT). The pilot areas have been periodically tested by colorimeter, contact sponge method, and bioluminescence, a diagram of the activities performed being shown in Table 38.1. In particular, set T0 includes the tests carried out before cleaning. Measures labeled T1–7 and T1–30 are reference values measured, respectively, 7 and 30 days after the treatment. The latter delay was chosen because it matches the complete curing time of the water-repellent product, i.e., the time taken by the oligomer organosilanes to completely polymerize and form the transpirant, repellent film on the protected surface. Sets T2 and T3 are carried out 1 and 2 years after the treatment, to log the performance of treatment over time. Moreover, the WSN has been recording the previously mentioned environmental parameters every 30 min during the whole duration of the activity. Figure 38.7 shows the results of the absorption tests. The water-repelling treatment performs better and lasts longer on the Doryphoros than on the Warrior. Again, this is to be ascribed to the coarser grain of the latter, whose greater porosity favors the penetration of the treatment, which is therefore less performant on the surface. Table 38.1 Schematic report of the performed measurements T0 T1–7 T1–30 T2 T3

Water absorption test X X X X X

ATP test X X X

Colour measures X X X X X

1132

C. Riminesi et al.

Fig. 38.7 Results of the absorption test

The colorimetric tests were performed considering the three CIELab color coordinates; nevertheless, parameters a* and b* have shown a negligible drift and are therefore considered not significant for assessing degradation. On the contrary, parameter L* showed a meaningful variation, which is reported in Fig. 38.8 (lower values correspond to darker colors). “As is” areas (XX) did not manifest an appreciable variation of luminance, which is stable around 45 (black lines). CX and CT areas show a higher index immediately after the mechanical cleaning (over 60, light grey), confirming its efficacy. Interestingly, the behavior is different between the areas facing east and those facing west. On the west-facing surfaces, the CT area shows a constant L* value during the 3 years (the efficacy of protective product is maintained practically constant from a chromatic point of view), while the CX area shows a lower value during the initial phase which then lowers, the darkening being caused by the growth of biological patina. On the east-facing surfaces, two opposite behaviors are shown: on the Doryphoros, the two areas are stable in time. On the contrary, on the Warrior the CT area features a stable value of luminosity, whereas the CX area shows a constant darkening due to the ongoing re-colonization by microorganisms. This may be explained by the higher surface roughness of this statue, which leads to a faster biological growth. Overall, the gradual loss of luminosity is to be ascribed to the exposure to the green environment, which is rich in biological aggressors and moisture. Water sources enabling colonizations are identified in rainfall and condensation

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1133

Fig. 38.8 Results of the colorimeter measures: trend of the L* (luminance) parameter

phenomena (Fig. 38.9); moreover, solar radiance also affects this drift, since UVA radiation induces on the long run a chemical variation of the treatment, impairing its performance and weakening the barrier effect towards surface moisture. Biodeteriorating activity has first been evaluated after the cleaning by bio-luminometer as indicated in Table 38.1. The amount of light produced, RLU, is directly proportional to the amount of microorganisms present on the surface through ATP concentration. Results in Table 38.2 are referred to unit area (RLU/cm2). The checked areas on both statues face east.

38.4

Conclusions

Defining and implementing maintenance and restoration practices is of utmost importance for the correct preservation of architectural heritage. In this work, guidelines are suggested for what concerns the conservation of statues and stone furniture in historical gardens, with a specific focus on marble artefacts. The protocol for preventive conservation is based on an initial condition of clean surface which follows a “timid intervention” A 3-year-long monitoring of two statues chosen as case studies in the garden of Villa Guicciardini Corsi Salviati has highlighted the following:

1134

C. Riminesi et al.

Fig. 38.9 Surface temperatures on the Doryphoros (red) in comparison with dew point (black) Table 38.2 ATP test results

Doryphoros

Warrior

Area XX CX CT XX CX CT

T1–7 [RLU/cm2] 3,2 2,1 0,6 2,8 1,8 0,8

T2 [RLU/cm2] 2,9 2,6 2,8 2,2 2,8 2,5

T3 [RLU/cm2] 2,5 2,5 2,6 3,0 2,7 3,1

• Efficacy of the water-repelling product after 2 years from the “timid intervention” is confirmed: the conservation state is manifestly different between the treated (CT) and non-treated (CX) areas. Biological colonization gradually resumes during the first 2 years on the CX areas, whereas such trend is much slower on CT areas (see Figs. 38.7 and 38.8).

38

Towards Preventive Conservation of Stone Artefacts in Historical Gardens. . .

1135

• Degradation rate, as determined by non-destructive techniques (water absorption test, colorimetric and ATP measurements), has been found higher on stones with greater roughness and porosity (Warrior statue). • A new cleaning seems suitable after about 4 years in order to restore the initial conditions. Some steps are required to properly plan maintenance of stone artefacts affected by biodegradation, in particular, it is fundamental to the following: • Assess the initial conservation state by means of a dedicated diagnostic campaign. • Identify the critical parameters and their relationship to the decay phenomena. • Monitor the selected parameters with non-destructive techniques to assess the effective risk for conservation. The techniques employed should feature low environmental impacts, low invasiveness, low energy consumption and low cost, and their reliability should be verified by spot tests performed in situ [5]. • Perform an etiological risk assessment over time. This step is mandatory to properly plan the intervention, be it a simple surface cleaning of a structural remediation. Focusing on the selected case study, cleaning of the marble surfaces by timid intervention should be repeated every 3 years on the non-treated areas (CX) and every 4 years on the treated areas. Such pattern can effectively prevent the biological recolonization of the artefacts which could impair the aesthetical comprehension of the artefacts. Acknowledgments The authors are thankful to the owners of Villa Guicciardini Corsi Salviati for their contribution. A special mention goes to the architect Tommaso Castellani (trustee) for his support during the measurement campaigns.

References 1. Manganelli Del Fá R, Casciani A, Vettori S, Cuzman OA, Tiano P, Rosa P, Riminesi C (in press) Sustainable conservation and restoration of historical gardens. Springer Handbook of Cultural Heritage Analysis 2. Vettori S, Cabassi J, Cantisani E, Riminesi C (2019) Environmental impact assessment on the stone decay in the archaeological site of Hierapolis (Denizli, Turkey). Science of the Total Environment 650:2962–2973 3. Mascalchi M, Orsini C, Pinna D, Salvadori B, Siano S, Riminesi C (2020) Assessment of different methods for the removal of biofilms and lichens on gravestones of the English Cemetery in Florence. Int Biodeterior Biodegradation 154:105041 4. Favero-Longo SE, Matteucci E, Pinna D, Ruggiero MG, Riminesi C (2021) Efficacy of the environmentally sustainable microwave heating compared to biocide applications in the devitalization of phototrophic communities colonizing rock engravings of Valle Camonica, UNESCO world heritage site, Italy. Int Biodeterior Biodegradation 165:105327 5. Bracci S, Cagnini A, Colombini MP, Cuzman OA, Fratini F, Galeotti M, Magrini D, Manganelli del Fà R, Porcinai S, Rescic S, Riminesi C, Salvadori B, Santagostino Barbone A, Tiano P (2016) A multi-analytical approach to monitor three outdoor contemporary artworks at the Gori collection (Fattoria di Celle, Santomato, Pistoia, Italy). Microchem J 124:384 878–888

1136

C. Riminesi et al.

6. Bracci S, Cagnini A, Colombini MP, Cuzman OA, Fratini F, Galeotti M, Magrini D, Manganelli del Fá R, Porcinai S, Rescic S, Riminesi C, Salvadori B, Santagostino Barbone A, Tiano P (2016) A multi-analytical approach to monitor three outdoor contemporary artworks at the Gori Collection (Fattoria di Celle, Santomato, Pistoia, Italy). Microchem J 124:878–888 7. Riggio M, Macchioni N, Riminesi C (2017) Structural health assessment of historical timber structures combining non-destructive techniques: the roof of Giotto’s bell tower in Florence. Struct Control Health Monitor 24(7):e1935 8. Mecchi AM, Sansonetti A, Realini M, Poli T (2008) A proposal for a common approach in choosing tests for the protocol evaluation of cleaning methods. In: Proceedings of the 11th international congress on deterioration and conservation of stone. Torun, Poland, pp 425–433 9. Activity in the framework of the project Smart4CH: Smart monitoring for cultural heritage, co-funded by Tuscany Region (POR FSE 2014–2020 - Fostering opportunities for human capital development through high-level training in multi-disciplinary areas) 10. Eastman C, Teicholz P, Sacks R, Liston K (2011) BIM handbook: a guide to building information modeling for owners, managers, designers, engineers and contractors, 2nd edn. Wiley, Hoboken 11. A. Guida, V. Porcari. 2018. Prevention, monitoring and conservation for a smart management of the cultural heritage. Int J Heritage Archit, 2(1), pp. 71–80 12. UNI 11432:2011, Cultural heritage – natural and artificial stone – determination of the water absorption by contact sponge 13. Tiano P, Pardini C (2004) Valutazione in situ dei trattamenti protettivi per il materiale lapideo. Proposta di una nuova semplice metodologia Arkos 5:30–36 14. Vandevoorde D, Pamplona M, Schalm O, Vanhellemont Y, Cnudde V, Verhaeven E (2009) Contact sponge method: Performance of a promising tool for measuring the initial water absorption. J Cult Heritage 10(1):41–47 15. Marques SM, da Silva JCGE (2009) Firefly bioluminescence: a mechanistic approach of luciferase catalyzed reactions. Life 61(1):6–17 16. EN 15886:2010 Conservation of cultural property – test methods – colour measurement of surfaces 17. Technical Report, Colorimetry 4th Edition, CIE 015:2018, International Commission on Illumination

Mechanical Monolithic Inertial Sensors for Historical and Archeological Heritage Real-Time Broadband Monitoring

39

Fabrizio Barone and Gerardo Giordano

Contents 39.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.1.1 Methodology and Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.1.2 Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.1.3 Folded Pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2 Folded Pendulum Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.1 Generalized Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.2 Extended Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3 The UNISA Folded Pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.4 Dynamic Behavior Assessment of the Trajan Arch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1138 1138 1139 1141 1143 1144 1149 1152 1160 1163

Abstract

The evaluation of the health state of historical and archeological infrastructures requires an analysis of their dynamical behavior in connection with natural and anthropic actions, whose accuracy is mainly determined by the band and sensitivity of the sensors. The latter are generally accelerometers, consisting of classical mechanical oscillators with a force feedback control, whose improvement has been directed and determined by the progress of the control electronics in the last decades. The mechanics, instead, has experienced only a limited evolution, mainly due to the introduction of new materials and modern machining techniques. But recently the introduction of the UNISA folded pendulum technological platform, a synthesis of more than 10 years of research and development, is allowing the implementation of very low-frequency compact monolithic oscillators ( f ¼ 1 α1 EJ rffiffiffiffiffiffiffiffi > < 1 f 2π L2 m 1 αi EJ α fi ¼ ) rffiffiffiffiffiffiffiffi ) 2 ¼ 2 ¼ const, 2 2π L m f α1 > 1 > 1 α2 EJ > : f2 ¼ 2 2π L m

ð42:1Þ

where αi are coefficients depending on the boundary conditions, L denotes the length of the cantilever beam, E is the elastic modulus, J is the moment of inertia of the cross section, and m is the distributed mass for unit length. Taking into account Eq. (42.1), the parametric analyses were first aimed at reproducing the experimental ratio between the bending frequencies along E-W and N-S directions. Figure 42.8 reports the sensitivity of the ratio between the second and the first bending frequencies in E-W direction vs. the height of the fixed restraint. No significant variations were observed due to the presence of constraints in the Y-direction (Hy). Figure 42.8 shows also the sensitivity of the same ratio in N-S direction. In this case, a little difference was observable by varying the height of the restraints in the Y-direction (Hy). Anyway, since in both cases the experimental ratio between the corresponding frequencies was not reached, additional analyses were performed by also varying the value of the elastic modulus of the internal filling. Results reported in Fig. 42.9 confirm that no variability on f2/f1 is produced by the mechanical properties of the filling material because of the linear dependence of the frequencies on the elastic properties selected in the model. The differences between numerical and experimental results are probably due to the experimental estimation of the second frequency that might be subject to measurement errors. To overcome this drawback, the FE model was identified to reproduce the first two main flexural frequencies. Figure 42.10 reports the results of the parametric analyses with the elastic modulus of the internal filling material varying between 1,000 and 2,500 N/mm2. The best numerical estimation of the dynamic behavior of TG was obtained by assuming a value of EN equal to 1,600 N/mm2, with fixed constraints in both directions up to a height of 19.30 m. Figure 42.11 shows the first three mode shapes obtained with the identified FE model together with the corresponding values of the numerical frequencies; the experimental values are quite well matched (see Table 42.1).

1234

G. Bartoli et al. Ratio f2/f1 (East-West) 4.9645

13.5

4.8098

H (m)

18

x

9

4.655

4.5

0

4.5002

0

1

2

3

4

5

6

7

8 9 10 11 12 13 14 15 16 17 18 Hy (m)

4.3455

Ratio f2/f1 (North-South) 18

5.3223

5.227

H (m)

13.5

5.1317

4.5

5.0364

x

9

0

0

1

2

3

4

5

6

7

8 9 10 11 12 13 14 15 16 17 18 Hy (m)

4.9411

Fig. 42.8 Variability of the ratio f2/f1 (second and first flexural frequency in the E-W direction) vs. the height of the restraints (up). Variability of the ratio f2/f1 (second and first flexural frequency in the N-S direction) vs. the height of the restraints (down)

42.5

Seismic Risk Assessment

The seismic risk of the masonry towers was evaluated according to the provisions of the “Italian Guidelines for the assessment and mitigation of the seismic risk of the cultural heritage” (IG) [44]. This document represents an innovative tool in the European context and proposes a methodology based on three levels of investigation:

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

42

1235

Ratio f2/f1 (East-West) 2500

4.9633

N

E (N/mm2)

4.8065

4.6497 1600 4.4929

1000

0

1

2

3

4

5

6

7

8 9 10 11 12 13 14 15 16 17 18 19 Hx = Hy (m)

4.3361

Ratio f2/f1 (North-South) 5.153

2500

N

E (N/mm2)

5.1017

5.0504 1600 4.9991

1000

0

1

2

3

4

5

6

7

8 9 10 11 12 13 14 15 16 17 18 19 Hx = Hy (m)

4.9478

Fig. 42.9 Variability of the ratio f2/f1 (second and first flexural frequency in the E-W direction) vs. the elastic modulus EN of the infill material and the height of the restraints (up). Variability of the ratio f2/f1 (second and first flexural frequency in the N-S direction) vs. the elastic modulus EN of the infill material and the height of the restraints (down)

• Level 1 (LV1) is a territorial risk level, in which the input is represented by the macro-seismic intensity and the vulnerability is evaluated according to a qualitative knowledge of the relevant structural parameters. The safety indexes are based on typological studies, related to the kind of the building (palace, church, tower), at a territorial scale. • Level 2 (LV2) is a local mechanical risk level, in which the response spectrum represents the input and the vulnerability is evaluated by analyzing the activation of collapse mechanisms in single parts of the structure (macro-elements). The

1236

G. Bartoli et al.

0.8512

4.5

0.8786

0.7382

0

4.5

9 13.5 Hy (m)

18

x

4.5

H (m)

0.9952

1.2906

0.9201

4.5

0.7966

0

1.2851

9

1.1485

4.5

0

4.5

9 13.5 Hy (m)

18

x

1.3306

0.943

0

4.5

9 13.5 Hy (m)

18

f2 num. [En=2500 N/mm2] 18

1.4216

13.5

0

H (m)

4.5

f1 num. [En=2500 N/mm2] 18

0.762

1.0722

1.0436

18

18

9

9

9 13.5 Hy (m)

9 13.5 Hy (m)

1.2014

1.1671

4.5

4.5

13.5

13.5

0

0

f2 num. [En=1600 N/mm2] 18

0.8138

1.4642

13.5

1.3192

9

1.1742

1.0119

4.5

1.0291

0.8754

0

x

x

H (m)

9

0.9642

0

x

1.1118

9

0

1.2284

13.5

1.0772

f1 num. [En=1600 N/mm2] 18

H (m)

1.1902

13.5

0

f2 num. [En=1000 N/mm2] 18

H (m)

x

H (m)

f1 num. [En=1000 N/mm2] 18

0

4.5

9 13.5 Hy (m)

18

0.8841

Fig. 42.10 Variability of the first (E-W) and the second (N-S) frequency vs. the elastic modulus EN of the infill material and of the height of the restraints

evaluation of the safety indexes still requires a few geometrical and mechanical parameters. • Level 3 (LV3) is a global mechanical risk level, in which the response spectrum represents the input and the vulnerability is evaluated performing nonlinear pushover analyses (capacity diagrams) or time-history analyses. The model asks

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

f1= 1.29 Hz

f2= 1.33 Hz

1237

f3= 5.184 Hz

Fig. 42.11 Main frequencies and corresponding modal shapes from FE results

for a detailed analysis of each single building, considered as a whole or as an assembly of macro-elements. The LV1 is a simplified expeditious analytical approach, which schematizes the tower as a cantilever beam with fixed base. The tower is assumed as constituted by n elements (sectors) each one with constant geometry, inertia, and material

1238

G. Bartoli et al.

properties. The LV2 consists of evaluating the local response of the structure by using collapse mechanisms method. The LV3 is aimed at assessing the global response of the tower by using, for instance, the FE method and the pushover analysis. If compared with the two previous levels, the last is the most accurate but, depending on the employed numerical approach and the type of analysis, it requires a large amount of input data together with a high computational effort. LV1 and LV3 levels are aimed at assessing, even though with different levels of refinement, the tower global structural behavior. To express the results, two synthetic safety indexes are evaluated with respect to the Life Safety Limit State (SLV). The first index, named seismic safety index Is,SLV, is the ratio between the return period TSLV of the seismic action which brings the tower to the SLV and the expected return time of the earthquake corresponding to the SLV (usually, as a reference, it is assumed TR,SLV ¼ 475 years): I s,SLV ¼

T SLV : T R,SLV

ð42:2Þ

A seismic safety index greater than one corresponds to a safe state for the tower, while a safety index lower than one highlights possible critical issues which ask additional in-depth investigations. The second index, the so-called acceleration factor fa,SLV, is the ratio between the peak ground acceleration aSLV that brings the tower to the SLV and the reference acceleration ag,SLV for the SLV (both are referred to a rigid soil condition): f a,SLV ¼

aSLV : ag,SLV

ð42:3Þ

The acceleration factor considers only one parameter of the seismic response spectrum; anyway it has the advantage of providing a quantitative indication of any deficiency in terms of mechanical strength of the structural system, as fa,SLV is a purely mechanical parameter. The index Is,SLV is based on the return periods of seismic demand and capacity, so it provides a direct evaluation of the possible vulnerability of the tower over time. Being not possible to identify in advance the most critical section, the two indexes must be evaluated according to the two principal directions for each section of the tower. In addition, when the tower is not an isolated structure, the effects introduced by the interaction with its adjacent lower buildings must be evaluated also. For the sake of simplicity, in several cases, the towers are analyzed as isolated structures even if they are connected with shorter neighboring buildings. While the effect of the confining buildings is currently assessed from a modal point of view and considered by a set of elastic constraints (like in [24]), few attention has been deserved to their effect on the assessment of the seismic risk. In this respect, it is worth remembering the recent contribution of de Silva et al. [76], Castellazzi et al. [77], Bartoli et al. [78, 79], and Torelli et al. [80]. The methodology proposed by the IG was applied in the next section to assess the seismic risk

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1239

of some of the towers illustrated in Figs. 42.4 and 42.5. In order to assess the seismic risk, their interaction with the adjacent buildings was comparatively assessed by performing parametric analyses.

42.5.1 Seismic Risk Assessment by the LV1 Approach The LV1 approach is specifically aimed at providing a comparative ranking risk and to highlight the need for subsequent in-depth investigations. The procedure consists of schematizing the tower as a cantilever beam subject to a system of static horizontal forces and is based on a limited number of geometrical and mechanical parameters that can be obtained by a qualitative visual inspections. According to the IG, the collapse of the tower can occur for a combined compressive and bending stress mode, i.e., a rocking-flexural mode. This assumption is open to criticisms since there are real cases in which the towers reveal completely different collapse modes (see, f.i. the shear collapse mode of the tower of Finale Emilia after the main shock of 2012, [3]). However, it should be observed that the purpose of this territorial risk level approach is mainly to provide comparative assessment and is not an absolute measure of the seismic risk. From the operational point of view, the tower is divided in n sectors by horizontal cuts. Each sector has the same geometrical, inertial, and material properties. The selection of the sectors should consider (i) the position of the openings, (ii) the levels of detachment of the tower from the neighboring buildings, (iii) the levels at which there is a masonry wall thickness reduction, and (iv) the levels where there are changes of materials and changes in the construction phases [77, 78, 79]. Once the sectors have been identified, at the base section of each sector the acting bending moment (seismic demand) and the resisting one (seismic capacity) are evaluated. The comparison of seismic demand and seismic capacity allows estimating the seismic safety of the structure. According to the Italian Code [44, 81, 82], the resisting bending moment Mui in the i-th section is given by: Mui ¼

  σ 0i Ai σ 0i Ai bi  : 2 0:85ai f d

ð42:4Þ

In Eq. (42.4) σ 0i denotes the average compressive stress due to the gravity loads, fd denotes the design compressive strength of the masonry, and Ai, bi, and ai, are geometrical parameters, i.e., the area of the cross section, the dimensions of the parallel, and perpendicular sides of the section (without openings) with respect to the seismic direction analyzed. The acting bending moment is evaluated by assuming a linear distribution of horizontal forces along the height of the tower. The force Fi to be applied at the center of mass of the i-th sector is given by:

1240

G. Bartoli et al.

W i zi W Fi ¼ P 0:85Sd ðT 1 Þ , n g W k zk

ð42:5Þ

k¼1

where Wi denotes the weight of the sector, zi is the height of its center of mass with respect to the foundations, and Sd(T1) is the ordinate of the design response spectrum, function of the main period T1 of the tower in the considered load direction. This value can be calculated by dividing the ordinate of the elastic response spectrum by the behavior factor q (assumed equal to 2.8 as suggested by [44] for structures with stiffness irregularities along the height); g denotes the gravity acceleration and W is the tower weight. Once the value of the seismic force in each sector is calculated, it is possible to evaluate the seismic resultant acting at the base section of the i-th sector by the following: n P j¼i

Fhi ¼ P n

k¼1

W jz j W k zk

0:85Sd ðT 1 Þ

W : g

ð42:6Þ

The acting bending moment Msi is then: n P

  Msi ¼ Fhi zhi  zi ,

zhi ¼

j¼i n P j¼i

W j z2j , W jz j

ð42:7Þ

where zhi denotes the quote of application of Fhi and zi* is the quote of the i-th section with respect to the basement. The acting bending moment Msi is function of the main period of the tower T1. According to [81] this period can be approximately calculated as (H expressed in m, T1 in s): T 1 ¼ 0:05H0:75 :

ð42:8Þ

Equation (42.8), as showed by Rainieri and Fabbrocino [83], overestimates T1 for values less than 1 s, while underestimated T1 for values greater than 1 s. The Authors proposed the following empirical correlation (used here together with Eq. (42.8)): T 1 ¼ 0:013H1:10 :

ð42:9Þ

Like the empirical correlation provided by the Italian Code, Eq. (42.9) provides the main period as function of the height H of the tower only. For comparative purposes, the main period of the tower can be also estimated employing the classical formula of a linear elastic cantilever beam with uniform cross section:

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

rffiffiffiffiffiffiffiffi γA T 1 ¼ 1:787H , EJg 2

1241

ð42:10Þ

where A is the cross section area, γ is the specific weight, E is the modulus of elasticity, and J denotes the cross section moment of inertia along the considered load direction. While Eqs. (42.8) and (42.9) express the main period of the tower as a function of its total height H only, Eq. (42.10) accounts for the geometric characteristics of the cross section and allows to estimate a different period for each loading direction. To account for the behavior of the structure at the ultimate limit state, i.e., to consider the nonlinear phenomena that occur as a result of the increasing damage induced by the seismic loading, the estimated values of T1 can be amplified by a factor of 1.4. The above procedure was applied to the following towers: • The Coppi-Campatelli tower (CC, Fig. 42.12) is a typological example of the constructive prototype called “casa torre” (tower house) which started to spread in San Gimignano during the twelfth–thirteenth century. Today the tower is included in an architectural complex that develops along via San Giovanni, the medieval entrance in San Gimignano. The whole complex is composed of three main buildings: (i) the CC tower house in Pisan style with two fornices (i.e., vaulted openings firstly introduced in the Italian town of Pisa), subsequently raised to obtain the present tower; (ii) the building on its right (Fig. 42.12), still in Pisan style, with a single vaulted arch; and (iii) a main building on its left, constituted by a portico with four openings in stone blocks on the ground floor and another lower level with vaulted ceilings in the basement. Two more buildings were added on the back façade. The main dimensions of the CC are as follows. The cross section of the tower is almost trapezoidal, with the two equal sides of about 6.6 m and the other two of about 7.8 m (South side) and 8.3 m (North side). The thickness of the walls ranges between 1.5 m at the base and 1.0 m at the upper level. With respect to via San Giovanni (the main street), the height of the tower is about 27.6 m; on the opposite side, due to the slope of the hill, the height is about 33.0 m. The latter is the height of the tower with respect to the foundation level. Overall, CC is internally divided in six levels, plus an underground one. The ground floor, the first floor, and the second floor are made up of wooden elements. The third floor is a barrel vault (probably the first intervention during the raising of the tower). The fourth, fifth, and sixth levels are wooden walkways that, through a series of wooden stairs, allow the access to the last level (a wooden floor). From a structural point of view, some specific features of the tower worth to be highlighted are the following: (i) a tilt of its axis of about 0.7–0.8 m along the southern direction, (ii) the presence at the ground levels of two fornices, and (iii) the incorporation of the lower levels, for about half of its height, into the aggregated buildings along via San Giovanni [78]. • The Cugnanesi tower (CU, Fig. 42.12), which is located in the hearth of the city center, is also included in a major architectonical complex: at the lower levels the

1242

G. Bartoli et al.

Fig. 42.12 The four investigated historic towers: Coppi-Campatelli (CC), Cugnanesi (CU), Becci (BE), and Chigi (CH)

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1243

tower is connected with two adjacent masonry buildings along its North side. Its height is about 42.8 m. The basement is a compact square parallelepiped made of stones, with a side of about 7.6 m and a height of about 5.2 m. On a lateral surface of the basement, along the North side, a little hole was dug. The inspection of such hole allowed to verify that the inner part of the parallelepiped is composed of well-constipated conglomerate (medium-size stones with good-quality mortar). Four lateral sustaining walls start above this block and gradually decrease their thickness from about 2.4 m at the base until 1.9 m at the top. A visual analysis of the tiny cavities of the walls along the height showed a multi-leaf structure. The walls are constituted by an internal core of heterogeneous stone blocks tied by a good mortar and two external cavernous limestone masonry layers. Seven light timber floors are along the height of tower, whose influence on the global structural behavior can be neglected. The last level is composed of a concrete slab sustained by a barrel vault. No crack pattern was visually detected on the structure [80]. • The Becci tower (BE, Fig. 42.12) dates back to the thirteenth century and overlooks the central “Piazza della Cisterna” (tank square). As the previous towers, also BE at the lower levels is incorporated in adjacent buildings built in a later period (with the exception of the South side). This tower is characterized by a satisfactorily regular geometry, since the most part of the few openings has limited dimensions. Exceptions to this regularity are due to large-size opening created in correspondence of some zones which connect the tower with the adjacent buildings. The external dimensions of the tower, together with the thickness of the wall, were measured by a detailed survey: the overall height is about 38 m and the plant has a slight taper to the upper levels. At the ground level the section sizes are as follows: 6.7 m on the North side, 6.8 m on the East side, 6.6 m on the South side, and 6.9 m on the West side. These dimensions at the last level reduce as follows: 6.2 m, 6.3 m, 6.2 m, and 6.4 m, respectively. Walls were made of multi-leaf masonry (two stone external faces and an inner core). The external stone layer has a thickness of about 40 cm while the internal one of about 25 cm. The inner core has an average thickness of about 1.6 m. • The Chigi tower (CH, Fig. 42.12). A specific feature that differentiates the CH tower from the other is the use of two different materials along the height. The lower part was built with stones, while the upper one was built with bricks. Like the previous tower, CH is nowadays incorporated (along the North, South, and East sides) into a palace facing the main square of San Gimignano, few meters far from other towers (the Rognosa, the Salvucci, and the Torre Grossa ones). The entire complex is the result of changes and modifications that occurred over the centuries: first of all the construction of the confining buildings and later the reshaping of the walls to extend the internal space. The tower was built in the second half of the thirteenth century, as it is confirmed by the presence of a sophisticate façade with openings of a different shape, such as single-lancet windows and segmental arches. About 50 years after its construction, the palaces on the South and East sides and later the one on the North side were built. As a result, today the tower is surrounded by masonry buildings up to

1244

G. Bartoli et al.

different levels (about 17 m on the South side and about 14 m on the North and East sides). Only the main façade, on the West side, is unconstrained, facing the main square. The tower has a square cross section with an external side of about 5.5  5.5 m and its height is about 27 m. The multilayer masonry walls have a thickness ranging from 1.6 m at the base up 1.3 m at the upper level. The lower walls have both external and internal faces composed of 30 cm-thick stone masonry (mainly travertine). The upper walls (from 13 m on) are composed of internal and external faces of bricks with a thickness of 25 cm. The internal filling, of unknown mechanical properties, is composed of heterogeneous material (remainder bricks tied by a poor mortar), and it appears, where it was possible to investigate, coherent. The dimensions of the prismatic blocks used for the masonry walls are about 50  30  30 cm for the stones and 12.5  30  5.5 cm for the bricks. The mortar joint thickness is minimal, thus producing good mechanical properties. The internal floors are constituted by masonry vaults, except for the fourth, the seventh, and the eighth floors which are made by wooden slabs. The provisions of the Italian Code [82] were used to estimate the mechanical properties of the masonry walls of the towers in absence of experimental results. In particular, four typologies of masonry textures were individuated in the four towers: (i) type B, uncut stone masonry with facing walls of limited thickness and infill core; (ii) type D, soft stone masonry (tuff, limestone, etc.); (iii) type E, dressed rectangular stone masonry; and (iv) type F, full brick masonry with lime mortar (used for characterizing the upper part of CH tower only). The reference intervals for the mechanical properties provided by the [82] were selected according to the characteristics of the masonry typologies existing in the Italian territory. These values refer to masonry with mortar of poor mechanical characteristics, and some correction factors were introduced accounting for increasing mechanical characteristics (to account, for instance, for good-quality mortar and thin joints) and decreasing ones (thick or poor internal core). When a limited level of knowledge is reached (KL1 according to [81]), the minimum strengths (uniaxial compressive strength fm and characteristic shear strength τ0) and the average values for the elastic moduli (Young modulus E and shear modulus G) have to be used (Table 42.3). Since the towers are largely incorporated into the neighboring buildings at the lower levels (Fig. 42.12), in order to apply the LV1 approach, three different schemes were considered: • Model A: the towers were analyzed as isolated constructions, i.e., without considering the presence of the neighboring buildings (it is implicitly assumed that the action offered by the neighboring structures is ineffective or that it can be lost during severe earthquakes). • Models B and C: the towers were still assumed as isolated constructions, but the tower height was assumed equal to the part of the structure emerging from the surrounding buildings (depending on the different height of such buildings,

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1245

Table 42.3 Mechanical properties according to [82] ( fm ¼ compressive strength, τ0 ¼ shear strength, E ¼ longitudinal modulus of elasticity) and correction factors corresponding to thick or poor internal core, good-quality mortar, and thin joints Mechanical characteristics Type of masonry B D E F

fm (N/mm2) 2.00 3.00 1.40 2.40 6.00 8.00 2.40 4.00

τ0 (N/mm2) 0.035 0.051 0.028 0.042 0.090 0.012 0.060 0.092

E (N/mm2) 1,020 1,440 900 1,260 2,400 3,200 1,200 1,800

Correction factors Thick or poor internal core 0.8

Goodquality mortar 1.4

Thin joints 1.2

0.9

1.5

1.5

0.7

1.2

1.2

0.7

1.5

1.5

different models B and C were considered accounting for the different emerging tower heights). The above models aimed at introducing lower and upper bounds. The presence of adjacent lower constructions can significantly alter the actual behavior of the towers that can be supposed to lie within the obtained range. In fact, the confining buildings reduce the effective slenderness, thus reducing the period. However, the surrounding buildings constitute stiffeners which might introduce localized areas of possible stress concentration and pounding. The estimated main periods of the four towers for the three models A, B, and C are summarized in Table 42.4. Results of the LV1 analyses are expressed in terms of acceleration factor fa,SLV, seismic safety index Is,SLV, and return period of the action causing the tower collapse TSLV in Table 42.5 (Chigi tower, CH), Table 42.6 (Coppi-Campatelli tower, CC), Table 42.7 (Cugnanesi tower, CU), and Table 42.8 (Becci tower, BE). Due to the low variation of the geometrical characteristics along the height, in all the analyzed cases the minimum ratio between the strength capacity and the seismic demand was obtained at the base section of each considered model (A, B, or C). For each case the minimum and the maximum T 1 calculated by Eqs. (42.8), (42.9), and (42.10) were considered. The LV1 analyses do not showed critical situations (although, in the case of the Chigi tower, the A model with main period evaluated according to Eq. (42.9) without the amplification factor 1.40 provided fa,SLV ¼ 0.97 and Is,SLV ¼ 0.91, case not reported in the Table). Overall it is possible to observe a general consistency of the results in terms of safety indexes. The smaller values of both fa,SLV and Is,SLV were obtained with the A models. It must however be observed that the approach adopted by the IG employs an equivalence with a cantilever masonry beam, where the failure mode considered is only the formation of a flexural hinge at the base. Failure modes

(42.8) (42.9) (42.10) (42.8) (42.9) (42.10) (42.8) (42.9) (42.10)

A

T1* ¼ 1.4  T1

C

B

Eq.

Model

10.5

13.4

CH H (m) 26.9

T1 (s) 0.59 0.49 0.63 0.35 0.22 0.16 0.28 0.17 0.08

T1* (s) 0.83 0.68 0.88 0.48 0.31 0.22 0.40 0.23 0.11 13.4

27.6

CC H (m) 33.1 T1 (s) 0.69 0.61 1.09 0.60 0.49 0.89 0.35 0.23 0.75

T1* (s) 0.97 0.85 1.52 0.84 0.69 1.05 0.49 0.32 0.20

Table 42.4 Main periods of the towers (empirical correlations and analytical expression)

15.5

17.4

BE H (m) 39.4 T1 (s) 0.79 0.74 1.33 0.43 0.29 0.28 0.39 0.26 0.22

T1* (s) 1.10 1.04 1.86 0.60 0.41 0.39 0.55 0.36 0.31

26.1

27.8

CU H (m) 42.8

T1 (s) 0.84 0.81 1.23 0.61 0.50 0.66 0.58 0.47 0.58

T1* (s) 1.17 1.13 1.72 0.85 0.70 0.93 0.81 0.65 0.82

1246 G. Bartoli et al.

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

Table 42.5 LV1 safety indexes of the CH tower

CH Model A

Direction N-S (X) E-W (Y)

Model B

N-S (X) E-W (Y)

Model C

N-S (X) E-W (Y)

Table 42.6 LV1 safety indexes of the CC tower

CC Model A

Direction N-S (X) E-W (Y)

Model B

E-W (Y)

Model C

N-S (X) E-W (Y)

Table 42.7 LV1 safety indexes of the CU tower

CU Model A

Direction N-S (X) E-W (Y)

Model B

N-S (X) E-W (Y)

Model C

N-S (X) E-W (Y)

T1* (s) 0.68 0.88 0.68 0.85 0.22 0.48 0.21 0.48 0.11 0.40 0.11 0.40

T1* (s) 0.85 1.52 0.85 1.25 0.69 0.87 0.25 0.32 0.20 0.32

T1* (s) 1.13 1.95 1.13 1.99 0.70 0.85 0.70 0.85 0.65 0.81 0.65 0.81

1247

fa,SLV () 1.01 1.29 1.08 1.38 >1.60 >1.60 >1.60 >1.60 >1.60 >1.60 >1.60 >1.60

Is,SLV () 1.03 2.29 1.30 2.95 >5.21 >5.21 >5.21 >5.21 >5.21 >5.21 >5.21 >5.21

TSLV (years) 487 1089 618 1399 >2475 >2475 >2475 >2475 >2475 >2475 >2475 >2475

fa,SLV () 1.38 > 1.60 1.23 > 1.60 1.31 1.58 1.85 1.85 > 1.60 > 1.60

Is,SLV () 1.58 > 5.21 1.09 > 5.21 1.34 2.51 4.54 4.54 > 5.21 > 5.21

TSLV (years) 749 > 2475 517 > 2475 637 1195 2158 2159 > 2475 > 2475

fa,SLV () 1.34 >1.60 1.34 >1.60 1.57 >1.60 1.57 >1.60 1.56 >1.60 1.56 >1.60

Is,SLV () 2.66 >5.21 2.66 >5.21 4.81 >5.21 4.81 >5.21 4.74 >5.21 4.74 >5.21

TSLV (years) 1260 >2475 1260 >2475 2283 >2475 2283 >2475 2251 >2475 2251 >2475

1248

G. Bartoli et al.

Table 42.8 LV1 safety indexes of the BE tower

BE Model A

T1* (s) 1.04 1.86 1.04 1.94 0.39 0.60 0.40 0.60 0.31 0.55 0.32 0.55

Direction N-S (X) E-W (Y)

Model B

N-S (X) E-W (Y)

Model C

N-S (X) E-W (Y)

Model A Direction North-South

30

fa,SLV () 1.07 >1.60 1.03 >1.60 0.77 1.36 0.96 1.40 0.96 1.36 0.98 1.39

Is,SLV () 1.26 >5.21 1.10 >5.21 0.91 2.85 0.89 3.17 0.89 2.86 0.96 3.10

Model A Direction East-West

30

Mu

Mu

Ms Tmax-UCr

25

Ms Tmax-UCr

25

Ms Tmin-UCr

15

Ms Tmin-Cr

15

10

10

5

5

0

0

0.5

1

1.5

M (kNm)

2

Ms Tmax-Cr

20

Ms Tmin-Cr

z (m)

z (m)

Ms Tmin-UCr

Ms Tmax-Cr

20

TSLV (years) 600 >2475 521 >2475 368 1355 425 1507 425 1359 454 1473

2.5 104

0

0

0.5

1

1.5

M (kNm)

2

2.5 104

Fig. 42.13 Ultimate moment Mu and acting moment Ms vs. the quote of the Chigi tower (model A, Tmax-UCr ¼ maximum period for uncracked sections, etc.)

due to shear or local collapse of the upper levels of the towers due to the presence of irregularities, high openings and bell towers, cannot be taken into account by the IG procedure. Figure 42.13 shows the results obtained for the model A of the CH tower considering the maximum and the minimum values of the main period evaluated by Eqs. (42.8), (42.9), and (42.10) in hypothesis of uncracked (UCr) and cracked (Cr) material. The capacity of the structure Mu and the seismic demand Ms was obtained as envelops along the two main seismic directions N-S and E-W which are represented along the height z of the tower. In all the sections it is always Mu > Ms (an exception occurs at the base for the N-S direction) denoting that the tower is in a safe condition with respect to the seismic hazard provided by the Italian Code. Similar conclusions were obtained for the models B and C. To evaluate the indexes Is,SLV and fa,SLV, the main period T1 was estimated by using the empirical and theoretical Eqs. (42.8), (42.9), and (42.10) considering the minimum and the

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1249

maximum values obtained with the formulae. All models provided indexes larger than one with the exception of the model A in the direction N-S where Is,SLV ¼ 0.91 and fa,SLV ¼ 0.97. This fact confirms the structural weakness of the ground floor, due to huge openings created at this level during the last century.

42.5.2 Seismic Risk Assessment by the LV3 Approach The global seismic response of the towers was investigated performing pushover analyses through FE models of the whole structure [44, 81, 82]. Focusing on the CH tower, four basic different models were developed. Each FE model accurately reproduces the geometry of the structure, including the internal masonry vaults but excluding the wooden slabs. The models TI-Om and TI-Nu consider the tower as an isolated structure (it is, hence, hypothesized that during an earthquake, the walls of the adjacent palaces are detached from the tower). The difference between the two models consists of the modeling strategy adopted for the masonry walls: in the first case (TI-Om) the mechanical properties of the masonry were homogenized along the thickness of the walls, and in the second case (TI-Nu), the internal and the external layers of the multi-leaf walls were modeled differentiating the mechanical properties with respect to the internal filling. The thickness of the layers was, as assessed according to a visual inspections of the masonry blocks, 25 cm for the bricks and 30 cm for the stones and was assumed constant along the height. In the third model, TC-Om, the masonry walls were considered as a homogeneous material, and the interaction with the adjacent buildings was taken into account by modeling an effective constraint only in the direction where the walls of the neighboring structures are compressed during the pushover analysis. The model TC-Nu is similar to TC-Om but separates the external layers and the internal filling.

Nonlinear Masonry Modeling To build the FE model of the CH tower, the macro-modeling approach was used with 8-node isoparametric finite elements having three degrees of freedom at each node (Solid 65). The nonlinear behavior of the masonry was reproduced by combining the plasticity model of Drucker-Prager (DP) and the William-Warnke (WW) smeared crack model. The Drucker-Prager Plasticity Model (DP) The Drucker-Prager plasticity criterion [84, 85] is typically used for pressuredependent inelastic materials such as soils, rocks, and concretes, and it is a modification of the Von Mises yield criterion that accounts for the hydrostatic stress component (the confinement pressure). The yield surface of the DP criterion depends on the first and the second invariant of the stress tensor and remains fixed in the stress space. Usually, the mean hydrostatic stress σ m and the effective shear stress σ are considered:

1250

G. Bartoli et al.

1 σ 2 ¼ sij sij , 2

σm ¼

σ ii , 3

ð42:11Þ

where sij are the deviatoric components of the stress tensor σ ij. The DP yield condition is defined as follows: F ¼ 3α σ m þ σ  k ¼ 0:

ð42:12Þ

The constants α and k are two parameters related to the friction angle φ and to the cohesion c of the material, according to the following equations: 2 sin φ α ¼ pffiffiffi , 3ð3  sin φÞ

6c cos φ k ¼ pffiffiffi : 3ð3  sin φÞ

ð42:13Þ

The two parameters α and k allow to evaluate the yield stresses in uniaxial tension and compression, ftDP and fcDP, respectively, by: f tDP ¼

p1ffiffi 3

k , þα

f cDP ¼

p1ffiffi 3

k : α

ð42:14Þ

In case of elastic-perfectly plastic behavior, the friction angle φ and the cohesion c are independent on the plastic deformation. The normal to the yield surface is calculated as follows: Q¼

@F 1 ¼ αδij þ sij : @σ 2σ

ð42:15Þ

The flow rule, that determines the direction of the plastic straining, is hence given by: fe_ gpl ¼ hλiP,

P ¼ βδij þ

1 s , 2σ ij

1 hλi ¼ ðλ þ jλjÞ, 2

ð42:16Þ

being P the plastic potential. If it is assumed α ¼ β (then P 5 Q), the flow rule is associated and the plastic straining occurs in direction normal to the yield surface. The experimental results available for soils and rocks show that the volumetric dilatation predicted by the associated DP flow rule is often larger than that obtained by the experiments. Therefore, a not associated flow rule should be used through a proper definition of the plastic potential. So, a third parameter is introduced, called the dilatancy angle δ. This parameter rules the flow degree of associativity. If δ ¼ φ the flow is associated, whereas if δ ¼ 0, no plastic volumetric strains will be produced. In conclusion, the definition of the DP model requires three parameters: the friction angle φ that describes the slope of the yield surface (if φ ¼ 0 there is no dependence on the hydrostatic pressure), the cohesion c (the yield strength at zero hydrostatic pressure), and the dilatancy angle δ. The DP yield surface can be considered as a smooth version of the Mohr-Coulomb yield surface, and usually the parameters c and φ are introduced in such a way that the circular cone of DP corresponds to the outer vertex of the hexagonal Mohr-Coulomb yield surface. The resulting surface is a right-circular cone

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1251

VIII

Fig. 42.14 Drucker-Prager yield surface in the HaighWestergaard stress space

[III

[II

VII

[I VI

pffiffiffi with apex at ρ ¼ k= 3α (Fig. 42.14). Depending on the parameter α and the ratio ftDP/fcDP, the yield function has three conical forms in plane stress: elliptic, parabolic, and hyperbolic. These forms can be analyzed in the 2D space (σ III ¼ 0), considering, the cross section of the DP cone in the plane (σ I, σ II). Taking into account that for masonry the ratio between the uniaxial compressive and tensile strengths is usually greater than 3, the conical form of the intersection is parabolic. The Willam-Warnke Smeared Crack Model (WW) The smeared crack model uses the failure surface proposed for concrete by Willam and Warnke (WW) [86, 87] implemented in ANSYS. According to this criterion, which sets a crushing and cracking rule, the elements are able to crack in tension and crush in compression. The WW surface shows an elliptic trace on the deviatoric sections in each sextant and a parabolic trace in the meridian sections. The definition of this surface requires to identify five parameters: the uniaxial compressive strength fcWW, the uniaxial tensile strength ftWW, the biaxial compressive strength fcb, and two parameters ρ1 and ρ2. These last define the curvature of the parabolic traces in the meridian sections for high values of the hydrostatic compression, for anomalies ζ ¼ 0 and ζ ¼ 60 . The failure surface has different expressions into the four domains: CCC, CCT, CTT, and TTT (C ¼ compression, T ¼ tension). In the CCC zone, f.i. this surface is expressed as follows: rffiffiffi 3 σ F ¼  1 ¼ 0, r ðσ m , ςÞ 5 f cWW 0

1

ð42:17Þ

where r and ζ denote the polar coordinates (radius and anomaly) in the deviatoric plane. The failure criterion, which accounts for cracking and crushing failure modes through a smeared approach, is completed by cut-off conditions. It allows cracking at each Gauss point in three orthogonal fixed directions, and cracking is modeled by modifying the material properties of the elements (a plane of weakness normal to the crack plane is introduced).

1252

G. Bartoli et al.

Despite the need pffiffiffifor five constants, in most practical cases, if the hydrostatic stress is limited to 3 fcWW, the failure surface can be identified by means of only the two parameters ftWW and fcWW. If this occurs, the remaining three parameters can be assumed as follows: f cb ¼ 1:2 f cWW ,

ρ1 ¼ 1:45 f cWW ,

ρ2 ¼ 1:725 f cWW :

ð42:18Þ

Two additional coefficients, denoted as βt and βc, are introduced to account for a shear strength reduction of the stress in order to produce sliding across the crack face for open (βt) or re-closed cracks (βc) [84]. If the WW failure criterion is joined with the DP plasticity criterion, the material behaves as an isotropic medium with plastic deformation, and cracking and crushing capabilities. In this case, the parameters required to identify the masonry nonlinear behavior are c, φ, and δ for DP and fcWW, ftWW, βt, and βc for WW. According to the experimental evidence, the combination of the WW and DP models must comply with the following criteria: (i) the tensile strength ftWW must be smaller than the tensile strength derived from the plasticity model ftDP and (ii) the compressive strength fcWW must be greater than the compressive strength derived from the plasticity model fcDP, to ensure the reproduction of plastic behavior of the masonry in the mixed tensile-compression zone. Both models have been extensively employed to model the inelastic behavior of the masonry material in the scientific literature. The DP model was used by Zucchini and Lourenço [88] discussing the homogenization approach for the simulation of the plastic deformation in masonry cells. The Authors showed that it is possible to account for the degradation of the masonry mechanical properties in compression. The DP criterion was also adopted by Cerioni et al. [89] to discuss the seismic behavior of the Parma Cathedral bell tower. Chiostrini et al. [90] combined DP and WW criteria to model the results of several diagonal tests on masonry panels, obtaining good agreement with the experimental results.

Pushover Analysis The four FE models of the Chigi tower were developed to perform nonlinear analyses by the criteria discussed in the previous section. The constitutive parameters used in input for DP and WW criteria are collected in Table 42.9. Then, only for models TI-Om and TC-Om, E, ftWW and fcDP were varied to obtain a realistic collection of structural responses (Table 42.10). FE Models of the Isolated Tower (TI-Om and TI-Nu) The FE models TI-Om and TI-Nu are showed in Fig. 42.15 (Fig. 42.16 illustrates the cases of confined tower). The models were built taking into account all the major openings, and the assumption of rigid ground foundation was done. The models were employed to perform pushover analyses by applying a uniform distribution of horizontal loads along the height, under constant gravity loads. The results are represented in terms of capacity diagrams in which the base shear V is plotted against

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1253

Table 42.9 Constitutive parameters used in the FE models of the Chigi tower Model

TI-Om TC-Om

TI-Nu TC-Nu

Stone masonry Brick masonry Stone masonry Brick masonry Stone infill Brick infill

DP criterion φ c (N/mm2) ( ) 0.115 40

δ ( ) 25

WW criterion fcWW ftWW (N/mm2) (N/mm2) 6.00 0.106

0.210

40

25

6.00

0.130

40

25

0.300

40

0.106 0.158

40 40

βc

βt

0.75

0.25

0.195

0.75

0.25

6.00

0.118

0.75

0.25

25

6.00

0.277

0.75

0.25

25 25

6.00 6.00

0.098 0.147

0.75 0.75

0.25 0.25

the displacement of the centroid of the upper level. Figure 42.17, as an example, reports these diagrams for model TI-Om varying the parameters as reported in Table 42.10. All the seismic directions (X and Y) were considered. The role of the elastic modulus E is evident since it, obviously, modifies the slope of the initial part of the diagrams. The maximum base shear varies from 0.15 W to 0.25 W, whereas a larger variability is obtained for the maximum displacement of the top of the tower [79]. A22 case reproduces the main typical feature of the masonry mechanical behavior, in particular the ratio between the compressive and the tensile strength of about 10. FE Models of the Confined Tower (TC-Om and TC-Nu) The FE models TC-Om and TC-Nu, which account for the surrounding structures, were based on the 3D models TI-Om and TI-Nu of the isolated tower by adding three couples of unitary length walls, placed along three sides of the tower (North, South, and East, Fig. 42.16). These unitary walls aim to simulate the level of constraint offered by the adjacent structures, and their stiffness was consequently evaluated in order to reproduce the stiffness of the confining walls. The evaluated equivalent stiffness, K¼Keq, represents a starting value, subsequently investigated by performing additional parametric analyses with respect to the level of constraint. The capacity diagrams obtained with the model TC-Om varying the parameters as reported in Table 42.10 are reported in Fig. 42.18. If compared with the results of the model TI-Om, an increase of strength and stiffness is clearly visible. The maximum base shear varies between 0.31 W and 0.80 W. To investigate the effectiveness of the confinement offered by the adjacent structures, the case A22 was selected together with the critical seismic direction (X, North, the one in which the base area with the major openings is subjected to tensile stresses). Results of three cases of the confined tower are compared in Fig. 42.19 with the isolated tower (TC-Om1 with K¼Keq/10, TC-Om2 with K¼Keq and TC-Om3 with K ¼ 10Keq). Overall, the pushover diagrams show that as the stiffness of the increases the top displacement decreases, even if base shear is higher. Consequently, the structural

Case A11 A12 A21 A22 B11 B12 B21 B22

Stone masonry ftWW (N/mm2) 0.106 0.106 0.106 0.106 0.212 0.212 0.212 0.212

fcDP (N/mm2) 0.493 0.493 0.986 0.986 0.986 0.986 1.973 1.973

E (N/mm2) 1,458 2,916 1,458 2,916 1,458 2,916 1,458 2,916

γm (kg/m3) 1,600 1,600 1,600 1,600 1,600 1,600 1,600 1,600

Brick masonry ftWW (N/mm2) 0.195 0.195 0.195 0.195 0.390 0.390 0.390 0.390

Table 42.10 Parametric investigation in the homogenized models of the Chigi tower (TI-Om and TC-Om) fcDP (N/mm2) 0.901 0.901 1.801 1.801 1.801 1.801 3.603 3.603

E (N/mm2) 2,363 4,726 2,363 4,726 2,363 4,726 2,363 4,726

γm (kg/m3) 1,800 1,800 1,800 1,800 1,800 1,800 1,800 1,800

1254 G. Bartoli et al.

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1255

Fig. 42.15 FE models of the isolated tower: (a) and (b) model TI-Om (masonry walls homogenized along the thickness); (c) model TI-Nu (masonry walls modeled as a multilayer material)

Fig. 42.16 FE models of the confined tower: (a) and (b) model TC-Om (masonry walls homogenized along the thickness); (c) model TC-Nu (masonry walls modeled as a multilayer material)

1256

G. Bartoli et al. Direction (-X) North

Direction (+X) South

0.25

0.25

0.2

0.2

0.15

0.15

A12 A21

0.1

A11

V/W

V/W

A11

A12 A21

0.1

A22

A22 B11

B11

B12

0.05

B12

0.05

B21

B21

B22 0

0

100

200

300

400

500

600

700

B22 0

800

0

100

200

Displacement (mm)

300

400

500

600

700

Direction (+Y) East

Direction ( -Y) West

0.25

0.25

0.2

0.2

0.15

0.15

A21

0.1

A11

V/W

V/W

A11 A12

A12 A21

0.1

A22

A22

B11 B12

0.05

B11 B12

0.05

B21

B21

B22 0

0

100

200

300

400

Displacement (mm)

800

Displacement (mm)

500

600

700

B22 0

0

100

200

300

400

500

600

700

Displacement (mm)

Fig. 42.17 FE model of the isolated tower TI-Om: capacity diagrams in the directions X and Y, varying the input parameters (Table 42.10)

behavior moves from a pseudo-ductile system to a brittle one that absorbs the seismic loads for its own strength. When the stiffness of the constraints increases, the cracking pattern varies from a diffused damage in the lower levels of the tower (TI-Om) to a concentrated damage close to the upper level of the constraining walls. In this case, the cracking pattern involves the higher levels of the tower, constituted by brick masonry with high mechanical properties. This fact leads to a structural response with higher base shear and a decreasing of the stretchiness. As a result, the presence of shorter adjacent structural elements considerably modify the seismic response of the tower. The walls of the tower along the seismic loading direction show diagonal cracks which start from the contact area with the confining buildings. This phenomenon presumes that the tower doesn’t collapse with a bending mode, and an important shear component influences the failure mechanism. This observation is confirmed by the analysis of the damage in the walls perpendicular to the seismic direction (North and South walls). These walls show a cracking pattern that moves from the lower part of the tower to the contact zone with the constraining walls.

Capacity Spectrum Method The safety checks were carried out through the capacity spectrum method (CSM), describing the capacity curve and the response spectrum in terms of spectral acceleration

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

Fig. 42.18 FE model of the confined tower TC-Om: capacity diagrams in the directions X and +Y, varying the input parameters (Table 42.10), for K ¼ Keq

1257

Confined model Direction (-X) North 0.45 0.4 0.35

V/W

0.3 A11

0.25

A12 0.2

A21 A22

0.15

B11 0.1

B12 B21

0.05

B22 0

0

10

20

30

40

50

60

70

80

Displacement (mm)

Confined model Direction (+X) South 0.9 0.8 0.7

V/W

0.6 A11

0.5

A12 0.4

A21 A22

0.3

B11 0.2

B12 B21

0.1

B22 0

0

20

40

60

80

100

120

140

Displacement (mm) Confined model Direction (+Y) East 0.7

0.6

0.5

0.4

V/W

A11 A12

0.3

A21

0.2

B11

0.1

B21

A22

B12

B22 0

0

20

40

60

80

100

120

Displacement (mm)

140

160

180

200

1258

G. Bartoli et al.

Fig. 42.19 FE model of the confined tower TC-Om with the input parameters A22 (Table 42.10), varying the degree of confinement

Parameters A22 Direction (-X) North

0.5

Isolated tower Confined tower Keq/10 Confined tower Keq Confined tower 10Keq

0.45 0.4 0.35

V/W

0.3 0.25 0.2 0.15 0.1 0.05 0 0

10

20

30

40

50

60

70

80

90

100

Displacement (mm)

and displacement in the acceleration-displacement response spectrum (ADRS) format. CSM compares the structure capacity (in the form of a pushover curve) with the demand (in the form of response spectrum) and provides an effective graphical evaluation of the seismic behavior of the structure. In fact, the intersection of the capacity spectrum with the demand spectrum identifies a point, denoted as performance point, which represents the condition where seismic capacity equals seismic demand. There are different versions of the CSM and many changes have been proposed to increase its efficiency. Each of them combines the pushover analysis of a multidegree-of-freedom (MDOF) system, represented by its capacity curve, with the analysis of the response spectrum of an equivalent single-degree-of-freedom system (SDOF). Herein, the seismic checks were performed with reference to the N2 method proposed by Fajfar [91, 92] that uses inelastic spectra. Fixed the viscous damping, Sae and Sde denote the acceleration and the displacement elastic spectra corresponding to the period T. The displacement spectrum is given by: Sde ¼

T2 S : 4π 2 ae

ð42:19Þ

The acceleration spectrum Sa and the displacement spectrum Sd of an inelastic SDOF system with a bilinear force-deformation relationship are determined as [93]: Sa ¼

Sae , Rμ

Sd ¼

μ μ T2 T2 Sde ¼ Sae ¼ μ 2 Sa , 2 Rμ Rμ 4π 4π

ð42:20Þ

where μ denotes the ductility factor (the ratio between the maximum displacement and the yield displacement) and Rμ is the reduction factor due to ductility, i.e., due to

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1259

the hysteretic energy dissipation of ductile structures. Rμ reduces the elastic force to account for the inelastic capacity of the structure. Several proposals have been made for the reduction factor, and the N2 employs the one reported in [93]; slightly improved by Miranda and Bertero [94]: R μ ¼ ð μ  1Þ Rμ ¼ μ

T þ1 Tc

T < Tc, T  Tc,

ð42:21Þ

where Tc is the characteristic period of the ground motion, i.e., the transition period where the constant acceleration segment of the response spectrum passes to the constant velocity segment. The above formula suggests that, in the medium- and long-period ranges, the equal displacement rule applies. Using Eqs. (42.19), (42.20), and (42.21), starting from the elastic design spectrum, the demand spectra (Sa and Sd) for the constant ductility factors μ in ADRS format can be obtained. The capacity diagram V  dc obtained with the pushover analysis is used to build that of the equivalent SDOF system. The displacement and the force of the SDOF system are evaluated by: F ¼

V , Γ

d ¼

dc , Γ

ð42:22Þ

where V is the base shear of the MDOF model and dc is the displacement of the control point. Γ is the modal participation factor that rules the transformation from the MDOF to the SDOF system both for displacements and forces. As a consequence, the force-displacement relationship determined for the MDOF system V  dc applies also to the equivalent SDOF system (in form of F*  d* diagram), provided that both force and displacement are divided by Γ. It is worth noting that the initial stiffness of the SDOF system remains the same as that defined by the base shear vs. the top displacement of the MDOF system. The modal participation factor is defined as follows: Γ¼

ΦT M τ , ΦT MΦ

ð42:23Þ

where Φ denotes the displacement shape of the structure, normalized to have dc equal to 1 (τ is the dragging vector and M the inertia matrix). The capacity diagram of the SDOF system is then idealized by a bilinear elasticperfectly plastic curve, with the procedure reported in the NTC-Instructions [82]. The elastic-perfectly plastic capacity is obtained imposing the equality of the areas under the two curves. The ultimate displacement of the bilinear system is given by the ultimate displacement of the capacity curve (with the condition that the decrease of the base shear force is less than 15% of the maximum shear). In conclusion, the elastic period T* and the yield strength Fy of the idealized bilinear system are:

1260

G. Bartoli et al.

rffiffiffiffiffiffi m T ¼ 2π , k qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffi    2      k d u  2k E , Fy ¼ k d u  

ð42:24Þ

where m* is the equivalent mass (m ¼ ΦTMτ), d u is the ultimate displacement of the equivalent system (corresponding to at least the 85% of the peak shear), E* is the energy dissipated by the equivalent system (equal to the area under the capacity curve), and k* is the initial stiffness of the equivalent bilinear system, defined as follows: k ¼

Fy : dy

ð42:25Þ

The capacity diagram of the equivalent SDOF system in the accelerationdisplacement (ADRS) format is obtained by dividing the forces in the forcedeformation (F*  d*) diagram by the equivalent mass m*: Sa ¼

F : m

ð42:26Þ

The seismic demand for the equivalent SDOF system is graphically defined by the intersection of the radial line corresponding to the elastic period of the idealized bilinear system T* with the elastic demand spectrum Sae. Such intersection defines the acceleration demand (strength) required for elastic behavior and the corresponding elastic displacement demand. The yield acceleration Sae represents both the acceleration demand and the capacity of the inelastic system. Different situations can be identified. (a) The elastic period T* is larger or equal to Tc, and the inelastic displacement demand Sd is equal to the elastic displacement demand Sde (rule of equal displacement). The ductility demand μ is equal to the reduction factor: du  dmax ¼ Sde ðT  Þ

T  Tc,

μ ¼ Rμ ,

ð42:27Þ

where dmax denotes the displacement demand. (b) The elastic period T* of the system is smaller than Tc, the ductility demand of the inelastic system d max is greater than the one of the elastic system of equal period de max , and the ductility displacement demand can be calculated as follows:  T  Sde ðT  Þ  1 þ Rμ  1 c Rμ T T Rμ ¼ ðμ  1Þ þ 1: Tc

du  dmax ¼

T < Tc,

ð42:28Þ

In both cases (T* < Tc and T*  Tc), the inelastic demand in terms of acceleration and displacement corresponds to the intersection point (performance

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1261

point) of the capacity diagram with the demand spectrum corresponding to the ductility demand μ. At this point, the ductility factor determined from the capacity diagram and the ductility factor associated with the intersecting demand spectrum are equal. The method compares the seismic capacity and the seismic demand in an effective graphical form using the acceleration-displacement response spectrum (ADRS). The CSM was first employed to investigate the safety of CH tower. The indexes Is,SLV and fa,SLV were calculated for the A22 case and for TI-Om and TC-Om models by considering three hypotheses of constraint (Table 42.11). Figure 42.20 shows the details of the CSM for the weakest seismic direction (X, North). In all the examined cases, the capacity of the structure is higher than the seismic demand, thus denoting safety conditions. This result is consistent with the one obtained by the LV1 analysis, where the only critical case was found at the base of the isolated tower for seismic loading acting in the N-S direction (Is,SLV ¼ 0.91). As the stiffness of the constraint increases, the seismic behavior of the tower changes. The seismic demand is lower for high values of the stiffness of the constraining walls than for the isolated tower or for low values of stiffness. The case TC-Om1 is the most critic for the structure; this is due to two reasons: the maximum level of damage is located in the lower part of the tower which is constituted by poor materials and the capacity of the system decreases as the stiffness of the constraining walls increases. For this reason the equivalent SDOF system has a higher yield value but a lower displacement capacity, showing the lower safety factor. In the other cases, the seismic capacity is lower but the strength and the stiffness of the system are higher, providing a lower seismic demand. Furthermore, the major stiffness of the complex leads to a local damage on the

Table 42.11 CH tower: results of LV3 analysis in terms of return period TSLV, seismic safety index Is,SLV and acceleration factor fa,SLV

South direction (+X)

North direction (X)

East direction (+Y)

West direction (Y)

Model TI-Om TC-Om1 TC-Om2 TC-Om3 TI-Om TC-Om1 TC-Om2 TC-Om3 TI-Om TC-Om1 TC-Om2 TC-Om3 TI-Om

T* (s) 1.10 0.76 0.39 0.24 0.65 0.69 0.42 0.25 0.79 0.70 0.42 0.27 0.98

fa,SLV () 5.01 2.99 2.60 3.83 1.22 1.23 1.38 2.82 2.82 2.15 2.18 4.19 3.69

TSLV (years) >2475 >2475 >2475 >2475 484 488 979 >2475 >2475 >2475 >2475 >2475 >2475

Is,SLV () >5.21 >5.21 >5.21 >5.21 1.02 1.03 2.04 >5.21 >5.21 >5.21 >5.21 >5.21 >5.21

1262

G. Bartoli et al. Parameters A22 ADRS (-X) North

Parameters A22 ADRS (-X) North

0.7

0.7

T =0.13 s B

0.6

T =0.13 s B

DS CS 0.6

T =0.39 s C

0.5

0.4

S a(g)

S a(g)

0.5

DS CS

T =0.39 s C

(a)

T*=0.65 s

0.4

(b)

T*=0.69 s

0.3

0.3 Fy*/m*

0.2 Fy*/m*

0.2

T =2.16 s

T =2.16 s D

D

0.1

0.1

dmax*

dmax*

du* 0 0

20

40

60

80

100

du*

0

120

0

20

40

S d (mm)

60

80

100

120

S d (mm)

Parameters A22 ADRS (-X) North

Parameters A22 ADRS (-X) North

0.7

0.7 TB =0.13 s

0.6

TB =0.13 s

DS CS

0.6

TC =0.39 s

DS (Elastic) DS (Inelastic) CS

TC =0.39 s T*=0.25 s

T*=0.42 s

0.5 Fy*/m* 0.4

0.4

(c)

S a(g)

S a(g)

0.5

Fy*/m*

0.3

0.2

(d) 0.3

0.2 T =2.16 s

T =2.16 s

D

D

0.1

0.1 dmax* du*

0 0

20

0 40

60

S d (mm)

80

100

120

0

dmax* 20 du*

40

60

80

100

120

S d (mm)

Fig. 42.20 Chigi tower, the capacity spectrum method (CSM) for the North direction (X) with different levels of constraint: (a) model of the isolated tower TI-Om; (b), (c), and (d) models of the confined tower with a stiffness of the constraining walls equal to Keq/10, Keq, and 10Keq

higher part of the tower involving the brick masonry with good mechanical properties. The same procedure was adopted to evaluate the two seismic indexes for the other three towers, and the results are reported in Tables 42.12, 42.13, and 42.14. For the Coppi-Campatelli tower, results do not highlight any critical situations (Table 42.12), providing safety indexes always higher than one, in agreement with those obtained with the LV1 approach. It is interesting to observe that the acceleration factor obtained with the LV3 model is always higher than those evaluated at LV1 (an exhaustive discussion of the CC seismic risk is reported in [78, 95]). Table 42.13 reports the safety indexes as obtained for the Cugnanesi tower. Also in this case, all the indexes are always higher than one and still in agreement with those obtained with the simplified LV1 approach. The obtained safety indexes of the Becci tower are reported in Table 42.14. The LV3 analyses have not highlighted critical situations (all the indexes values are higher than one), and also in this case a general agreement between LV3 and LV1 approach is observed. Overall, the LV3 analyses have not highlighted critical situations, and all the examined cases are in respect to the seismic checks as the ultimate capacity displacement of each system is always greater than the

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1263

Table 42.12 CC tower: LV3 safety indexes (IT¼isolated tower; CT¼confined tower) CC South direction (+X)

North direction (X)

East direction (+Y)

West direction (Y)

Model IT CT1 CT2 IT CT1 CT2 IT CT1 CT2 IT

T* (s) 1.54 0.84 0.51 1.34 0.89 0.72 1.07 0.84 0.47 0.91

fa,SLV () 2.21 4.40 7.63 3.08 4.33 4.38 2.87 1.33 1.57 1.79

TSLV (years) >2475 >2475 >2475 >2475 >2475 >2475 >2475 625 992 1519

Is,SLV () – – – – – – – 1.32 2.09 3.20

Table 42.13 CU tower: LV3 safety indexes (IT¼isolated tower; CT¼confined tower) CU South direction (Y) North direction (+Y) East direction (+X) West direction (X)

Model IT IT CT IT IT CT

T* (s) 1.15 1.32 0.82 1.25 1.26 0.93

fa,SLV () 1.52 4.22 4.06 4.27 4.04 2.91

TSLV (years) 921 >2475 >2475 >2475 >2475 >2475

Is,SLV () 1.94 – – – – –

Table 42.14 BE tower: LV3 safety indexes (IT¼isolated tower; CT¼confined tower) BE East direction (+X) West direction (X) North direction (+Y) South direction (Y)

Model IT CT IT CT IT CT IT

T* (s) 1.85 0.64 1.98 0.77 1.98 0.77 2.01

fa,SLV () 3.49 5.23 2.12 2.62 2.05 3.46 1.91

TSLV (years) >2475 >2475 >2475 >2475 2321 >2475 1894

Is,SLV () – – – – 4.89 – 3.99

corresponding displacement demand. The LV3 results confirm those obtained with the LV1 approach, and, in addition, the indexes are always greater than those obtained with the first level of evaluation (the LV1 approach is more conservative), denoting a general coherence of the two methods. Nevertheless, despite such a general coherence, a clear trend was not observed for the acceleration factor and the seismic vulnerability index within the analyzed bounds.

1264

G. Bartoli et al.

References 1. Bowitz E, Ibenholt K (2009) Economic impacts of cultural heritage – research and perspectives. J Cult Herit 10(1):1–8. https://doi.org/10.1016/j.culher.2008.09.002 2. Fioravanti M, Mecca S (eds) (2011) The safeguard of cultural heritage: a challenge from the past for the Europe of tomorrow. Firenze University Press 3. Acito M, Bocciarelli M, Chiesi C, Milani G (2014) Collapse of the clock tower in finale Emilia after the May 2012 Emilia Romagna earthquake sequence: numerical insight. Eng Struct 72:70– 91. https://doi.org/10.1016/j.engstruct.2014.04.026 4. Augenti N, Parisi F (2010) Learning from construction failures due to the 2009 L’Aquila, Italy, earthquake. J Perform Constr Facil 24:536–555. https://doi.org/10.1061/(ASCE)CF.1943-5509. 0000122 5. Casciati S, Al-Saleh R (2010) Dynamic behavior of a masonry civic belfry under operational conditions. Acta Mech 215(1–4):211–224. https://doi.org/10.1007/s00707-010-0343-4 6. Gentile C, Saisi A (2013) Operational modal testing of historic structures at different levels of excitation. Constr Build Mater 48:1273–1285. https://doi.org/10.1016/j.conbuildmat.2013.01.013 7. Ivorra S, Pallarés FJ (2006) Dynamic investigations on a masonry bell tower. Eng Struct 28(5): 660–667. https://doi.org/10.1016/j.engstruct.2005.09.019 8. Pieraccini M, Dei D, Betti M, Bartoli G, Tucci G, Guardini N (2014) Dynamic identification of historic masonry towers through an expeditious and no-contact approach: application to the “Torre del Mangia” in Siena (Italy). J Cult Herit 15:275–282. https://doi.org/10.1016/j.culher. 2013.07.006 9. Saisi A, Gentile C, Ruccolo A (2016) Pre-diagnostic prompt investigation and static monitoring of a historic bell-tower. Constr Build Mater 122:833–844. https://doi.org/10.1016/j.conbuildmat. 2016.04.016 10. Zini G, Betti M, Bartoli G (2022) A quality-based automated procedure for operational modal analysis. Mech Syst Signal Process 164:108173. https://doi.org/10.1016/j.ymssp.2021.108173 11. Bartoli G, Betti M, Spinelli P, Tordini B (2006) An “innovative” procedure for assessing the seismic capacity of historical tall buildings: the “Torre Grossa” masonry tower. In: Proceeding of the V international conference on structural analysis of historical constructions, SAHC 2006, New Delhi, November 6–8, 2006, pp 929–937 12. Bassoli E, Vincenzi L, D’Altri AM, de Miranda S, Forghieri M, Castellazzi G (2018) Ambient vibration-based finite element model updating of an earthquake-damaged masonry tower. Struct Control Health Monit 25:e2150. https://doi.org/10.1002/stc.2150 13. Casolo S (1998) A three-dimensional model for vulnerability analysis of slender medieval masonry towers. J Earthq Eng 2(4):487–512. https://doi.org/10.1080/13632469809350332 14. Casolo S, Milani G, Uva G, Alessandri C (2013) Comparative seismic vulnerability analysis on ten masonry towers in the coastal Po Valley in Italy. Eng Struct 49:465–490. https://doi.org/10. 1016/j.engstruct.2012.11.033 15. Lucchesi M, Pintucchi B (2007) A numerical model for non-linear dynamic analysis of slender masonry structures. Eur J Mech A Solids 26:88–105. https://doi.org/10.1016/j.euromechsol.2006. 02.005 16. Peña F, Lourenço PB, Mendes N, Oliveira DV (2010) Numerical models for the seismic assessment of an old masonry tower. Eng Struct 32:1466–1478. https://doi.org/10.1016/j. engstruct.2010.01.027 17. Sarhosis V, Milani G, Formisano A, Fabbrocino F (2018) Evaluation of different approaches for the estimation of the seismic vulnerability of masonry towers. Bull Earthq Eng 16(3):1511– 1545. https://doi.org/10.1007/s10518-017-0258-8 18. Ferrante A, Loverdos D, Clementi F, Milani A, Formisano A, Lenci S, Sarhosis V (2021) Discontinuous approaches for nonlinear dynamic analyses of an ancient masonry tower. Eng Struct 230: 111626. https://doi.org/10.1016/j.engstruct.2020.111626 19. Bartoli G, Betti M, Giordano S (2013) In situ static and dynamic investigations on the “Torre Grossa” masonry tower. Eng Struct 52:718–733. https://doi.org/10.1016/j.engstruct.2013.01.030

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1265

20. Ivorra S, Pallarés FJ, Adam JM (2009) Experimental and numerical results from the seismic study of a masonry bell tower. Adv Struct Eng 12:287–293. https://doi.org/10.1260/ 136943309788251641 21. Zonta D, Zanon P, Molinari M, Toffaletti S, Anese GA, Busetto M (2004) Aggiornamento della vulnerabilità sismica della Torre Civica di Portogruaro basato su misure vibrazionali 2004. In: Atti dell’XI Congresso Nazionale L’Ingegneria Sismica in Italia, Genova, Italia, 2004 (in Italian) 22. Standoli G, Salachoris GP, Masciotta MG, Clementi F (2021) Modal-based FE model updating via genetic algorithms: Exploiting artificial intelligence to build realistic numerical models of historical structures. Constr Build Mater 303:124393. https://doi.org/10.1016/j.conbuildmat. 2021.124393 23. Lacanna G, Betti M, Ripepe M and Bartoli G (2020) Dynamic Identification as a Tool to Constrain Numerical Models for Structural Analysis of Historical Buildings. Front. Built Environ. 6:40. https://doi.org/10.3389/fbuil.2020.00040 24. Gentile C, Saisi A (2007) Ambient vibration testing of historic masonry towers for structural identification and damage assessment. Constr Build Mater 21(6):1311–1321. https://doi.org/10. 1016/j.conbuildmat.2006.01.007 25. Saisi A, Gentile C (2015) Post-earthquake diagnostic investigation of a historic masonry tower. J Cult Herit 16(4):602–609. https://doi.org/10.1016/j.culher.2014.09.002 26. Cavalagli N, Comanducci G, Ubertini F (2018) Earthquake-induced damage detection in a monumental masonry bell-tower using long-term dynamic monitoring data. J Earthq Eng 22(1): 96–119. https://doi.org/10.1080/13632469.2017.1323048 27. Pierdicca A, Clementi F, Isidori D, Concettoni E, Cristalli C, Lenci S (2016) Numerical model upgrading of a historical masonry palace monitored with a wireless sensors network. Int J Masonry Res Innov 1:74–99. https://doi.org/10.1504/IJMRI.2016.074748 28. Ubertini F, Comanducci G, Cavalagli N, Pisello L, Materazzi AL, Cotana F (2017) Environmental effects on natural frequencies of the San Pietro bell tower in Perugia, Italy, and their removal for structural performance assessment. Mech Syst Signal Process 82:307–322. https:// doi.org/10.1016/j.ymssp.2016.05.025 29. Barsocchi P, Bartoli G, Betti M, Girardi M, Mammolito S, Pellegrini D, Zini G (2021) Wireless Sensor Networks for continuous Structural Health Monitoring of Historic Masonry Towers. Int J Archit Herit 15(1):22–44. https://doi.org/10.1080/15583058.2020.1719229 30. Clementi F, Pierdicca A, Formisano A, Catinari F, Lenci S (2017) Numerical model upgrading of a historical masonry building damaged during the 2016 Italian earthquakes: the case study of the Podestà palace in Montelupone (Italy). J Civ Struct Heal Monit 7(5):703–717. https://doi. org/10.1007/s13349-017-0253-4 31. Erdogan YS (2017) Discrete and continuous finite element models and their calibration via vibration and material tests for the seismic assessment of masonry structures. Int J Archit Herit 11(7):1026–1045. https://doi.org/10.1080/15583058.2017.1332255 32. Castellazzi G, de Miranda S, Mazzotti C. (2012) Finite element modelling tuned on experimental testing for the structural health assessment of an ancient masonry arch bridge. Math Prob Eng 495019. https://doi.org/10.1155/2012/495019 33. Pieraccini M, Betti M, Forcellini D, Dei D, Papi F, Bartoli G, Facchini L, Corazzi R, Cerisano KV (2017) Radar detection of pedestrian-induced vibrations of Michelangelo’s David. PLoS One 12(4):e0174480. https://doi.org/10.1371/journal.pone.0174480 34. Pieraccini M, Miccinesi L, Rojhani N (2017) A GBSAR operating in monostatic and bistatic modalities for retrieving the displacement vector. IEEE Geosci Remote Sens Lett 14(9):1494– 1498. https://doi.org/10.1109/LGRS.2017.2717857 35. Ramos LF, Aguilar R, Lourenco PB, Moreira S (2013) Dynamic structural health monitoring of Saint Torcato Church. Mech Syst Signal Process 35:1–15. https://doi.org/10.1016/j.ymssp. 2012.09.007 36. Aoki T, Sabia D, Rivella D (2008) Influence of experimental data and FE model on updating results of a brick chimney. Adv Eng Softw 39(4):327–335. https://doi.org/10.1016/j.advengsoft. 2007.01.005

1266

G. Bartoli et al.

37. Girardi M, Padovani C, Pellegrini D, Robol L (2019) A model updating procedure to enhance structural analysis in the FE code NOSA-ITACA. J Perform Constr Facil 33(4):04019041. https://doi.org/10.1061/(ASCE)CF.1943-5509.0001303 38. Pavlovic M,Trevisani S, Cecchi A. (2019). A procedure for the structural identification of masonry towers. J Nondestruct Eval; 38(2). DOI https://doi.org/10.1007/s10921-019-0575-8 39. Pellegrini D, Girardi M, Lourenço PB, Masciotta MG, Mendes N, Padovani C, Ramos LF (2018) Modal analysis of historical masonry structures: linear perturbation and software benchmarking. Constr Build Mater 189:1232–1250. https://doi.org/10.1016/j.conbuildmat. 2018.09.034 40. Torres W, Almazán JL, Sandoval C, Boroschek R (2017) Operational modal analysis and FE model updating of the Metropolitan Cathedral of Santiago, Chile. Eng Struct 143:169–188. https://doi.org/10.1016/j.engstruct.2017.04.008 41. Bru D, Ivorra S, Betti M, Adam JM, Bartoli G (2019) Parametric dynamic interaction assessment between bells and supporting slender masonry tower. Mech Syst Signal Process 129:235– 249. https://doi.org/10.1016/j.ymssp.2019.04.038 42. Diaferio M, Foti D, Giannoccaro NI, Ivorra S (2017) Model updating based on the dynamic identification of a baroque bell tower. Int J Safety Secur Eng 7(4):519–531. https://doi.org/10. 2495/SAFE-V7-N4-519-531 43. Pierdicca A, Clementi F, Fortunati A, Lenci S (2019) Tracking modal parameters evolution of a school building during retrofitting works. Bull Earthq Eng 17(2):1029–1052. https://doi.org/10. 1007/s10518-018-0483-9 44. DPCM (2011) Direttiva del Presidente del Consiglio dei Ministri per la valutazione e riduzione del rischio sismico del patrimonio culturale con riferimento alle NTC 2008. G. U. n. 47 del 26.02.2011 (in Italian) 45. Anzani A, Binda L, Carpinteri A, Invernizzi S, Lacidogna G (2010) A multilevel approach for the damage assessment of historic masonry towers. J Cult Herit 11(4):459–470. https://doi.org/ 10.1016/j.culher.2009.11.008 46. Binda L, Zanzi L, Lualdi M, Condoleo P (2005) The use of georadar to assess damage to a masonry bell tower in Cremona, Italy. NDT E Int 38(3):171–179. https://doi.org/10.1016/j. ndteint.2004.03.010 47. Ivorra S, Pallares FJ, Adam JM, Tomas T (2010) An evaluation of the incidence of soil subsidence on the dynamic behavior of a Gothic bell tower. Eng Struct 32(8):2318–2325. https://doi.org/10.1016/j.engstruct.2010.04.007 48. Russo G, Bergamo O, Damiani L, Lugato D (2010) Experimental analysis of the Saint Andrea Masonry Bell Tower in Venice. A new method for the determination of Tower Global Young’s Modulus E. Eng Struct 32(2):353–360. https://doi.org/10.1016/j.engstruct.2009.08.002 49. Ferrante A, Clementi F, Milani G (2019) Dynamic behavior of an inclined existing masonry tower in Italy. Front Built Environ 5:33. https://doi.org/10.3389/fbuil.2019.00033 50. Ramos LF, Marques L, Lourenço PB, De Roeck G, Campos-Costa A, Roque J (2010) Monitoring historical masonry structures with operational modal analysis: two case studies. Mech Syst Signal Process 24(5):1291–1305. https://doi.org/10.1016/j.ymssp.2010.01.011 51. Bernardeschi K, Padovani C, Pasquinelli G (2004) Numerical modelling of the structural behavior of Buti’s bell tower. J Cult Herit 5:371–378. https://doi.org/10.1016/j.culher.2004. 01.004 52. Carpinteri A, Invernizzi S, Lacidogna G (2006) Numerical assessment of three medieval masonry towers subjected to different loading conditions. Masonry Int 19:65–76 53. Girardi M, Padovani C, Pagni A, Pasquinelli G (2010) Numerical modeling of masonry towers: the case study of the Rognosa Tower in San Gimignano. In: Proceedings of the fourth international conference on structural engineering, mechanics and computation, SEMC 2010, Cape Town, September 6–8 54. Micelli F, Cascardi A (2019) Structural assessment and seismic analysis of a 14th century masonry tower. Eng Fail Anal. https://doi.org/10.1016/j.engfailanal.2019.104198 55. Facchini L, Betti M (2014) An efficient Bouc & Wen approach for seismic analysis of masonry tower. Fract Struct Integr 29:139–149. https://doi.org/10.3221/IGF-ESIS.29.13

42

Seismic Assessment of Historic Masonry Towers: Non-invasive Techniques and. . .

1267

56. Facchini L, Betti M (2016) Simplified seismic analysis of disordered masonry towers. ASCEASME J Risk Uncertainty Eng Syst, Part A: Civil Eng 2(2):C4015010. https://doi.org/10.1061/ AJRUA6.0000856 57. D’Ambrisi A, Mariani V, Mezzi M (2012) Seismic assessment of a historical masonry tower with nonlinear static and dynamic analyses tuned on ambient vibration tests. Eng Struct 36:210– 219. https://doi.org/10.1016/j.engstruct.2011.12.009 58. Milani G, Casolo S, Naliato A, Tralli A (2012) Seismic assessment of a medieval masonry tower in northern Italy by limit, non-linear static and full dynamic analyses. Int J Architect Herit 6(5): 489–524. https://doi.org/10.1080/15583058.2011.588987 59. Milani G, Shehu R, Valente M (2017) Seismic assessment of masonry towers by means of nonlinear static procedures. Proc Eng 199:266–271. https://doi.org/10.1016/j.proeng.2017. 09.022 60. Pieraccini M (2013) Monitoring of civil infrastructures by interferometric radar: a review. Sci World J 786961. https://doi.org/10.1155/2013/786961 61. Pieraccini M, Miccinesi L (2019) Ground-based radar interferometry: a bibliographic review. Remote Sens 11(9):1029. https://doi.org/10.3390/rs11091029 62. Pieraccini M, Dei D, Mecatti D (2013) Interferometric radar for testing large structures with a built-in seismic accelerometer. Sensors Actuators A Phys 204:25–30. https://doi.org/10.1016/j. sna.2013.09.028 63. Pieraccini M, Dei D, Mecatti D, Parrini F (2013) Dynamic testing of historic towers using an interferometric radar from an unstable measurement position. J Nondestruct Eval 32(4):398– 404. https://doi.org/10.1007/s10921-013-0193-9 64. Lee H, Lee JH, Kim KE, Sung NH, Cho SJ (2013) Development of a truck-mounted arc-scanning synthetic aperture radar. IEEE Trans Geosci Remote Sens 52(5):2773–2779 65. Pieraccini M, Miccinesi L (2016) ArcSAR: theory, simulations, and experimental verification. IEEE Trans Microw Theory Tech 65(1):293–301. https://doi.org/10.1109/TMTT.2016.2613926 66. Pieraccini M, Luzi G, Mecatti D, Fratini M, Noferini L, Carissimi L, Franchioni G, Atzeni C (2004) Remote sensing of building structural displacements using a microwave interferometer with imaging capability. NDT E Int 37(7):545–550. https://doi.org/10.1016/j.ndteint.2004.02.004 67. Hu C, Wang J, Tian W, Zeng T, Wang R (2017) Design and imaging of ground-based multipleinput multiple-output synthetic aperture radar (MIMO SAR) with non-collinear arrays. Sensors 17(3):E598. https://doi.org/10.3390/s17030598 68. Pieraccini M, Miccinesi L (2019) An interferometric MIMO radar for bridge monitoring. IEEE Geosci Remote Sens Lett 16(9):1383–1387. https://doi.org/10.1109/LGRS.2019.2900405 69. Tarchi D, Oliveri F, Sammartino PF (2012) MIMO radar and ground-based SAR imaging systems: equivalent approaches for remote sensing. IEEE Trans Geosci Remote Sens 51(1): 425–435. https://doi.org/10.1109/TGRS.2012.2199120 70. Zeng T, Mao C, Hu C, Yang X, Tian W (2015) Multi-static MIMO-SAR three dimensional deformation measurement system. In: Proceedings of the 2015 IEEE 5th Asia-Pacific Conference on Synthetic Aperture Radar (APSAR), Singapore, September 1–4, 2015, pp 297–301 71. Pieraccini M (2016) Extensive measurement campaign using interferometric radar. J Perform Constr Facil 31(3):04016113. https://doi.org/10.1061/(ASCE)CF.1943-5509.0000987 72. Lourenço PB, Rots JG, Blaauwendraad J (1995) Two approaches for the analysis of masonry structures: micro and macro-modeling. HERON 40(4):313–340 73. Theodossopoulos D, Sinha B (2013) A review of analytical methods in the current design processes and assessment of performance of masonry structures. Constr Build Mater 41:990– 1001. https://doi.org/10.1016/j.conbuildmat.2012.07.095 74. Zucchini A, Lourenço PB (2002) A micro-mechanical model for the homogenisation of masonry. Int J Solids Struct 39(3–4):3233–3255. https://doi.org/10.1016/j.ijsolstr.2008.09.034 75. Betti M, Galano L, Vignoli A (2016) Finite element modelling for seismic assessment of historic masonry buildings. In: D’Amico S (ed) Earthquakes and their impact on society. Springer Natural Hazards, pp 377–415. https://doi.org/10.1007/978-3-319-21753-6_14 76. de Silva F, Ceroni F, Sica S, Silvestri F (2017) Non-linear analysis of the Carmine bell tower under seismic actions accounting for soil–foundation–structure interaction. Bull Earthq Eng 16(7):2775–2808. https://doi.org/10.1007/s10518-017-0298-0

1268

G. Bartoli et al.

77. Castellazzi G, D’Altri AM, de Miranda S, Chiozzi A, Tralli A (2018) Numerical insights on the seismic behavior of a nonisolated historical masonry tower. Bull Earthq Eng 16(2):933–961. https://doi.org/10.1007/s10518-017-0231-6 78. Bartoli G, Betti M, Vignoli A (2016) A numerical study on seismic risk assessment of historic masonry towers: a case study in San Gimignano. Bull Earthq Eng 14:1475–1518. https://doi. org/10.1007/s10518-016-9892-9 79. Bartoli G, Betti M, Galano L, Zini G (2019) Numerical insights on the seismic risk of confined masonry towers. Eng Struct 180:713–727. https://doi.org/10.1016/j.engstruct.2018.10.001 80. Torelli G, D’Ayala D, Betti M, Bartoli G (2020) Analytical and numerical seismic assessment of heritage masonry towers. Bull Earthq Eng 18(3):969–1008. https://doi.org/10.1007/s10518019-00732-y 81. NTC (2008) D.M. del Ministero delle Infrastrutture e dei Trasporti del 14/01/2008. Nuove Norme Tecniche per le Costruzioni. G.U. n. 29 del 04.02.2008, S.O. n. 30 (in Italian) 82. NTC-Instruction (2009) Circolare n. 617 del 2 febbraio 2009. Istruzioni per l’Applicazione Nuove Norme Tecniche Costruzioni di cui al Decreto Ministeriale 14 gennaio 2008 (in Italian) 83. Rainieri C, Fabbrocino G (2011) Il periodo elastico delle torri in muratura: correlazioni empiriche per la previsione. In: Atti del XIV Congresso Nazionale L’Ingegneria Sismica in Italia, Bari, Italia, 2011 (in Italian) 84. ANSYS Inc (1998) ANSYS Manual. Southpoint 85. Drucker D, Prager W (1952) Soil mechanics and plastic analysis or limit design. Q Appl Math 10(2):157–165 86. Salari MR, Saeb S, Willam KJ, Patchet SJ, Carrasco RC (2004) A coupled elasto-plastic damage model for geo-materials. Comput Methods Appl Mech Eng 193(27–29):2625–2643. https://doi.org/10.1016/j.cma.2003.11.013 87. Willam KJ, Warnke ED (1975) Constitutive model for the triaxial behaviour of concrete. In: Proceeding of the international association for bridge and structural engineering, Bergamo, 1975 88. Zucchini A, Lourenço P (2007) Mechanics of masonry in compression: result s from a homogenisation approach. Comput Struct 85(3–4):193–204. https://doi.org/10.1016/j. compstruc.2006.08.054 89. Cerioni R, Brighenti R, Donida G (1995) Use of incompatible displacement modes in a finite element model to analyze the dynamic behavior of unreinforced masonry panels. Comput Struct 57(1):47–57. https://doi.org/10.1016/0045-7949(94)00590-Y 90. Chiostrini S, Galano L, Vignoli A (1998) In situ tests and numerical simulations on structural behaviour of ancient masonry. In: Proceedings of Monument-98, workshop on seismic performance of monuments, Lisbon, November 12–14, 1998 91. EUROCODE 8 (2002) Design of structures for earthquake resistance of structures. CEN, European Committee of Standardisation, Brussels 92. Fajfar P, Gaspersic P (1996) The N2 method for the seismic damage analysis of RC buildings. Earthq Eng Struct Dyn 25(1):31–46. https://doi.org/10.1002/(SICI)1096-9845(199601)25: 13.0.CO;2-V 93. Vidic T, Fajfar P, Fischinger M (1994) Consistent inelastic design spectra: strength and displacement. Earthq Eng Struct Dyn 23:502–521. https://doi.org/10.1002/eqe.4290230504 94. Miranda E, Bertero VV (1994) Evaluation of strength reduction factor for earthquake resistant design. Earthquake Spectra 10(2):357–379. https://doi.org/10.1193/1.1585778 95. Bartoli G, Betti M, Monchetti S (2017) Seismic risk assessment of historic masonry towers: comparison of four case studies. J Perform Constr Facil 31(5):04017039. https://doi.org/10. 1061/(ASCE)CF.1943-5509.0001039

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

43

Susanna Bracci, Maria Piera Caggia, Emma Cantisani, Tommaso Ismaelli, Massimo Limoncelli, Cristiano Riminesi, Giuseppe Scardozzi, and Silvia Vettori

Contents 43.1 43.2 43.3 43.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Natural and Anthropogenic Hazards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.1 Methods for the Reconstruction of the Ancient Landscape . . . . . . . . . . . . 43.4.2 Methods for the Reconstruction of the Geomorphological and Tectonic Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.3 Methods for the Study of the Relationships Among Environmental Conditions and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.4 Methods for the Study of Ancient Building Materials . . . . . . . . . . . . . . . . . . 43.4.5 Methods for the Reconstruction of the Ancient City Landscape . . . . . . . 43.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1270 1271 1272 1275 1275 1276 1277 1285 1287 1290 1292

Abstract

The paper focuses on the strong connections between natural resources, environment, and urban development in the Hellenistic, Roman, and Byzantine city of Hierapolis of Phrygia (Pamukkale, Turkey). The ancient city was founded on a travertine terrace crossed by an active fault, responsible for impressive geothermal phenomena, i.e., S. Bracci (*) · E. Cantisani (*) National Research Council, Florence, Italy e-mail: [email protected]; [email protected] M. P. Caggia (*) · T. Ismaelli (*) · G. Scardozzi (*) National Research Council, Lecce, Italy e-mail: [email protected]; [email protected]; [email protected] M. Limoncelli (*) University of Palermo, Palermo, Italy C. Riminesi (*) · S. Vettori (*) Institute of Heritage Science – CNR, National Research Council, Florence, Italy e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_43

1269

1270

S. Bracci et al.

flowing of thermal water, emission of gases, and frequent earthquakes, while the surrounding territory offered various stone qualities (travertine, alabaster, marbles, etc.). These environmental features affected the cultural identity of Hierapolis and its urban layout, conditioning also the construction techniques and the monumentalization of the city. In recent year, a multidisciplinary research has been performed, which saw the factual integration of archaeological, geo-archaeological, and archaeometric perspectives, in order to reconstruct how the natural resources and environmental phenomena impacted on the urban history of Hierapolis and the conservation of its monuments. The main methods for the reconstruction of the ancient landscape, the tectonic context, and the cityscape are discussed.

43.1

Introduction

Natural resources and environment are strictly connected to the historical development of ancient cities, both offering construction material to their inhabitants and influencing their building practices. Moreover, environmental conditions clearly affect the conservation of the architectural heritage. These statements are even more true in the case of an ancient city that lies in a highly seismic region, such as Hierapolis of Phrygia (Pamukkale, Turkey) (Fig. 43.1), which was founded on a travertine terrace crossed by an active fault responsible for impressive geothermal phenomena, i.e., flowing of thermal water, emission of gases, and frequent earthquakes. Because of this combination of peculiar natural features and ancient remains, the city of Hierapolis and its territory constitute an exceptional context listed by UNESCO as a protected site of worldwide importance. The archaeological area has been the subject of systematic investigations by the Italian Archaeological Mission since 1957. In the years 2013–2016 a multidisciplinary research, the Marmora Phrygiae project, has been performed integrating archaeological, geo-archaeological, and archaeometric perspectives in order to deeply understand the strong connections between the historic development of Hierapolis, the natural resources of the surrounding territory, and the conservation of the cultural heritage. The research units involved in the project were Institutes of the Italian National Research Council (IBAM, Istituto per i Beni Archeologici e Monumentali; IGAG, Istituto di Geologia Ambientale e Geoingegneria; ICVBC, Istituto per la Conservazione e la Valorizzazione dei Beni Culturali) and the University of the Salento (Dipartimento di Beni Culturali), also in cooperation with the Unitat d’Estudis Arqueométrics of the Institut Català d’Arqueologia Clàssica (UEA-ICAC, Tarragona). The Marmora Phrygiae project, supported by the integration of experts from a range of disciplines (archaeology, ancient topography, history of art, architecture, geology, geophysics, physics, chemistry, biology, geochemistry, remote sensing, information technology, electrical engineering, Roman law), allowed for a holistic knowledge of the ancient city, from the general features of the territory and urban area (reconstruction of the extractive districts of southern Phrygia, study of the urban layout and its relation with the environmental phenomena, reconstruction of the architectural heritage, and its visual perception) to more detailed aspects linked to building technology and material culture (study of architectural

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1271

Fig. 43.1 SRTM DEM of the Lykos valley and the Uzunpınar plateau: in red the ancient marble quarries

tradition, identification of workshops, characterization of mortars and plasters, polychromy in the sculpture, etc.). On this background, the paper focuses on the multidisciplinary approach of the project in order to provide suggestions for further research activities, conservation policies, and management of archaeological sites.

43.2

Study Area

Hierapolis was founded during the late third century BC [1, 2]. During the Roman Imperial time, the city was famous for the presence of the Sanctuary of Hades (Ploutonion) [3, 4], identified by the ancients as one of the entrances to the

1272

S. Bracci et al.

underworld. Its economy was based on the exploitation of the natural resources (such as marble and alabaster, agricultural products, and sheep-farming) and the production of purple-colored wool textiles. During the Byzantine period Hierapolis became a pilgrimage center because of the Tomb of the Apostle Philip. The city was abandoned in the 13th–14th century AD. The remains of Hierapolis lie on a travertine platform looking into the plain of the Lykos River, the current Çürüksu River [5] (Fig. 43.2), and was characterized by a regular urban layout based on a main route (plateia) [6–8]. During the Augustan and Julio-Claudian period, the main sanctuaries and the principal public buildings were monumentalized with a large use of marbles. Another important building phase followed the earthquake of AD 60, which strongly damaged Hierapolis: the urban area was expanded to the north and the south. In particular, during the HadrianicAntonine age, a very large Agora [9], together with a Theatre [10], was built in the new northern block. A significant building phase characterized also the Severan age, when some edifices were re-built (such as the Theatre [11] and Temple A in the Apollo Sanctuary [12]) and other new ones were constructed (such as the Nymphaeum of the Tritons [13]). The urban landscape dramatically changed during the early Byzantine period, when for the first time Hierapolis was enclosed by long city walls; in the same time, many efforts were devoted to monumentalize the area close to the Tomb of St. Philip thanks to the construction of the Martyrion and the Basilica dedicated to the Apostle [14].

43.3

Natural and Anthropogenic Hazards

From a structural point of view, the terrace of Hierapolis is bordered upstream by the Karahayıt segment of the Pamukkale fault zone, the northern master fault of the Denizli basin [15], which caused the lifting of the above Uzunpınar plateau, formed by metamorphic rocks (gneiss, schist, and marble). Significant travertine deposits lie along the entire outer edge of the terrace of Hierapolis and in the underlying terraced slope leading down toward the valley floor. Within the terrace of Hierapolis, the urban area is crossed by the so-called fault of the Ploutonion, which consists of a morphological sub-rectilinear step, at least 1.5 km long. Along this step are aligned various cracks from which gas and thermal water emerge today. They constitute the most evident effects of the fault activity and produce the famous white travertine incrustations covering the slopes descending to the valley below. The natural hazards linked to the presence of the seismic fault crossing the urban area of Hierapolis can be grouped in three types: (i) earthquakes, (ii) gas emissions, and (iii) flooding of calcareous waters. Moreover, (iv) the extreme climatic conditions of the site have to be added to these hazards. (i) Hierapolis is situated in one of the most tectonically active regions of Asia Minor, and over centuries it was struck by numerous earthquakes [1, 2, 16]

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

Fig. 43.2 Archaeological map of the urban area of Hierapolis

1273

1274

S. Bracci et al.

(Fig. 43.3a). Their effects are clearly recognizable in the numerous fissures crossing the ground surface. The geological instability and the periodical recurrence of disastrous events are also reflected in the numerous traces of ancient restorations detected during the archaeological excavations [17]. (ii) The gas emissions, which flood from the seismic fissures, are composed of a very high concentration of carbon dioxide, up to 96% in the natural cave of the Ploutonion [18, 19] (Fig. 43.3b). These emissions are characterized by high temperature and have a dramatic effect on the preservation on marble and travertine structures. Moreover, the deadly effect of gas vapors nowadays

Fig. 43.3 Nymphaeum of the Tritons, the rear wall collapsed after the earthquake of the seventh century AD (a); Ploutonion, hot water, and gas vapors in front of the grotto (b); Central Agora, submerged remains of the eastern portico (c); tourists swimming in the lake among the ancient remains of the Central Agora (d); the Theatre under the snow (e); an ancient alabaster quarry partially destroyed by modern extraction activities (f)

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1275

kills insects, birds, and mammals, strongly corroborating the ancient account on the use of these emissions to sacrifice bulls and birds by the ancient priests and inhabitants of Hierapolis. In fact, as recorded by ancient writers, the bulls were suffocated at the entrance of the grotto of the Ploutonion by vapors, causing the astonishment of the worshippers. (iii) Thermal waters (33–36  C) flood especially from the seismic fissure located close to the Ploutonion. They are calcium bicarbonate-sulfate and present high concentrations of CO2, which is responsible of the re-precipitation of carbonatic phases, and travertine encrustations [19, 20]. During the Hellenistic and Roman Imperial times, the thermal water was controlled and conducted through channels to the cultivated fields. After an earthquake occurred in the seventh century AD, a higher piezometric level made this control more difficult, and waters began to overflow in the lower sectors of the urban area. Some sectors of the city were completely submerged by waters, which produced thick (up to 4 m high) travertine incrustations over the monuments [21]. In other areas closest to the springs, such as the Central Agora (Fig. 43.3c), waters formed a natural lake still visible today [22]. (iv) Hierapolis has a mild climate with temperature rising to 40  C during the summer and falling to 5  C in the winter (Fig. 43.3e) [19]. This high thermal excursion determines micro-structural stress to the stone materials, which are accentuated in the areas close to the seismic fissures by the already mentioned warm emissions. In addition to these natural hazards, the ancient remains of Hierapolis are affected by anthropogenic phenomena of deterioration, which are largely increased during the last decades. Indeed, in the last years the peculiar natural landscape of Hierapolis attracted masses of tourists, and now the site is visited by 1.7 million tourists every year [23]. This high touristic pressure focuses on the Theatre and the lake submerging the porticoes of the Central Agora, where the bathing activity among the collapsed columns is permitted (Fig. 43.3d). Lastly, even the countryside is affected by anthropogenic hazards. These include the construction of infrastructures and accommodation facilities for the tourists, which produced the expansion of the urbanized areas [24]. Moreover, ancient farms are threatened by mechanized agricultural works and, in some cases, also by illegal digs. Finally, a massive resumption of extraction activities particularly affected several ancient quarries of travertine and alabaster (Fig. 43.3f).

43.4

Methods

43.4.1 Methods for the Reconstruction of the Ancient Landscape The research carried out in the territory of Hierapolis allowed for the reconstruction of the historical development of this area from the Prehistoric times to the Ottoman age. The research methodology was mainly based on the systematic archaeological survey, which was performed using GPS systems for the positioning of ancient

1276

S. Bracci et al.

features on the archaeological map. The field walking was supported by the integration of various methods and technologies: in particular, processing, analysis, and interpretation of multi-temporal space photos and high-resolution satellite images (from 1961 to 2016), and geophysical prospecting [25, 26]; moreover, in some cases aerial documentation was acquired using drones [27–30]. Optical satellite images were also processed for the production of ortho-images, space maps, and topographic maps used as base maps for the archaeological cartographies of the territory and the GIS platform devoted to the research project [31–33]. Also, remote sensing applications with high-resolution radar images taken by COSMO-SkyMed satellites were successfully tested by means of some Synthetic Aperture Radar images collected in Spotlight mode [34]. Moreover, 3D digital models of the territory, with large and medium resolution, were processed from both radar (Shuttle Radar Topography Mission) and optical satellite data (Advanced Spaceborne Thermal Emission and Reflectance Radiometer; Ikonos-2). The investigated area consists of the northern sector of the Lykos river valley, where Hierapolis lies, the Uzunpınar plateau, and the western sector of the Çal plateau, north of the city, corresponding to the ancient territory of Hierapolis. The investigations allowed for the discovering of many archaeological remains, dating to the centuries between the Hellenistic age and the Byzantine period: the aqueducts that brought drinking water to the city [35]; the ancient quarries of travertine, calcite alabaster, white and gray marble, and polychromatic breccia [36, 37]; the roads network [38]; rural villages and farms [39–41]; installations for olive oil and wine production [42]; funerary areas [43]; sanctuaries [44]; land divisions [45]; and palaeo-environmental features, such as a lake, today dried up, but mentioned in a second-century AD inscription [46]. The research of the years 2013–2016 has been particularly focused on the study of the marble and alabaster quarries of the territory of Hierapolis and the strategies of their exploitation: systematic topographical surveys and documentation of these quarries have been performed in order to determine the extent of the various extraction sectors, the techniques and strategies for extraction, the quantities of extracted material, the areas where the detritus was dumped, and the presence of activities linked to the roughing out of specific artifacts [47–49]. The geolocation of the quarrying areas, by means of high-precision differential GPS, was performed both on topographical maps (scale 1:25,000) and high-resolution satellite orthoimages (scale from 1:10,000 to 1:5000). In addition, the available multi-temporal satellite dataset made it possible the documentation of the transformations occurred in the territory during the last 15 years, allowing for the identification of quarrying sectors that had been destroyed or filled in with detritus.

43.4.2 Methods for the Reconstruction of the Geomorphological and Tectonic Context Thanks to systematic geo-archaeological surveys and geophysical investigations, a geological large-scale map of the urban area and surrounding territory was produced [50]. The geological investigation of the site of Hierapolis was also integrated with a

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1277

topographical survey of the traces of the seismic events detectable on the ground surface (i.e., fissures, wetlands, springs, lakes, sinkholes). These archaeoseismological features were georeferenced in the digital large-scale archaeological map of Hierapolis [51]. Their relationship with the archaeological remains allowed for determining a chronology of these features, which can be linked to the various earthquakes that damaged Hierapolis over the centuries. So, this geo-archaeological study enabled us to identify seismological features which existed before the foundation of the city (such as the fault of the Ploutonion) and numerous others that were produced by the earthquakes of the Imperial, Medieval, and Modern eras (Fig. 43.4d). With the aim to reconstruct the tectonic context in specific areas of the city and study the approaches to construction that were adopted in order to cope with the seismic nature of the geological substrate, also geophysical surveys were performed combining different techniques: Ground-Penetrating Radar (GPR), Electrical Resistivity Tomography (ERT), Magnetometry, and Seismic Refraction Tomography (SRT) [52]. A specific case study regarded the Sanctuary of Apollo, where geophysical investigations aimed to understand the connection between Temples A and C, and the underlying seismic fault [12, 53–55] (Fig. 43.4a, b). The ERT profiles documented the main fissure and other minor fractures parallel to the main one, filled with material, which was plausibly placed there in ancient times in order to level the second terrace of the sanctuary. The main tectonic fissure beneath Temple A was further investigated by means of georadar and seismic surveys to establish how the monument was built with respect to the geomorphological context. At about 4 m below the flooring of the cella of the temple, the seismic investigation documents the presence of the seismic fracture that crosses the bedrock below the eastern half of the podium. The cavity seems to be broader and deeper in the area of the cella (Fig. 43.4c), precisely below the hole for libations visible in the limestone floor (bothros) [56]. The northward continuation of the fissure, along the edge of the terrace, led to the construction, in the Julio-Claudian period, of a large barrel vault made of travertine blocks, which rested on the edge of the bedrock and supported the building itself and the steps leading to the terrace above.

43.4.3 Methods for the Study of the Relationships Among Environmental Conditions and Materials The archaeological site of Hierapolis offered an exceptional possibility to study the relationship between environmental conditions and the state of conservation of natural materials, such as marbles and travertines, and artificial materials, such as mortars and plasters. Environmental factors (i.e., water contact through thermal springs, high thermal stress, CO2 degassing) and geological aspects (e.g., presence of seismic faults, frequent earthquakes, formation of travertine) play a fundamental role in the durability and conservation state of the stone materials. Preliminary IR thermography survey was performed to investigate pieces of land, the buildings, and ruins affected by thermal anomalies due to the direct exposure to

1278

S. Bracci et al.

Fig. 43.4 Sanctuary of Apollo: ERT profiles georeferenced in the archaeological map (a–b; hypothetical position of the fissures in the bedrock indicated by green lines); Temple A, seismic time slice at 4 m depth (from the cella pavement) georeferenced on the monument plan (c); large fissure produced by an earthquake south of the sacred area (d)

gases discharged by fractures and hot springs [57]. These thermal anomalies are due to the presence of a subterranean flux of warm fluids connected to the fractures. Environmental parameters (i.e., temperature (T), relative humidity (RH), CO2, H2S, CO, O2) were monitored in several buildings near the faults. In addition, the geochemical composition of thermal waters and dissolved gases were investigated in different areas in order to evaluate their role in the conservation of the exposed materials. Moreover, a microsampling was carried out to characterize and evaluate the various weathering phenomena [18]. The IR survey was useful to choose the area of interest for the monitoring of environmental parameters (T and RH) and the concentration of carbon dioxide (CO2)

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1279

with respect to the diagnostic spot analysis (physical-chemical investigations). The monitoring aimed to understand and register the daily, weekly, and seasonal variations of the following parameters: • T and RH in several areas of the archaeological site (Theatre, Sanctuary of Apollo, and Ploutonion), close to the thermal spring waters or to the active fractures • Surface temperature (Tsup) of stone blocks directly exposed to solar radiation and to thermal spring waters (Sanctuary of Apollo and Ploutonion) • T, RH, and CO2 concentration inside the grotto of the Ploutonion The monitoring of the environmental parameters (T, Tsup, and RH) was performed by commercial dataloggers (Tinytag Plus 2, Gemini Data Logger – http://www.geminidataloggers.com/data-loggers/tinytag-plus-2/tgp-4500), designed to detect the temperature and relative humidity in adverse environmental conditions both outdoor and indoor. The dataloggers, closed into strong waterproof cases (IP65 Standard), are efficient in terms of battery life and storing data. These features make them particularly suitable for the environmental monitoring in the harvest condition of the archaeological site of Hierapolis. In order to measure the gases (CO2, O2, H2S, and CO) concentration close to the fractures, inside the grotto of the Ploutonion – thus in presence of spring waters where the RH is over 95% – specific dataloggers were realized by integrating industrial gas sensors in a customized electronic board, also considering that the high level of CO2, the necessity of big storing data capability (acquisition every 15 min for 1 year), the use of batteries without energy harvesting systems (e.g., solar panels), and the hazardous environmental condition for electronic circuits cannot allow the use of commercial products. For this purpose, the gas sensors by Alphasense Ltd (http://www.alphasense.com), whose technical characteristics are shown in Table 43.1, were chosen. In general, the gas sensors must be installed together with temperature and relative humidity sensors in order to check the environmental constrains (Table 43.1) and to compensate the sensors’ response in relation to changes in T and RH. The low cross sensitivity and long-term stability are the features that allowed the success of these customized sensors. The CO2-D1 gas sensor is a potentiometric electrochemical gas sensor that responds to CO2 as a gas ion selective electrode. The potential in output to the sensor was measured using high impedance circuitry. For this purpose, the conditioning unit made from Alphasense Ltd was used together with a datalogger to safely register the sensor output (Fig. 43.5a). The CO2 concentration values were obtained by interpolation on the mastercurve for the registered temperature at the same time of the sampling of the tension in output. The interpolation rule gives the concentration in % with the sampled mV in input. The CO2 sensor and the other gases sensors were set up without pumped aspiration in order to avoid the use of plug-in, thus taking advantage of the natural diffusion of gases. The sensors were assembled in an IP-65 case to protect the electronic control unit and the battery from water condensation and acid attack.

Measured size CO2

O2

H2S

CO

Type CO2-D1

O2-A2

H2S-AE

CO-AE

Up to 1000 ppm

Up to 2000 ppm

0%–30%

Range 0.2%–95%

10–25 nA/ppm

24 months

Operating life >24 months

1280 S. Bracci et al.

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1281

Fig. 43.5 Setup of the CO2 datalogger (a) and its typical calibration curve (b); setup of the O2, CO, and H2S sensors inside the IP-65 case (c); positioning of the dataloggers nos. 1–2 in the Ploutonion (d); the datalogger no. 1 in a niche along the northern entrance (e); the datalogger no. 2 inside the natural cave (f)

The gases sensors for O2, H2S, and CO were assembled together in a different case (Fig. 43.5c) because the supply and data sampling have to comply at different constraints with respect to CO2 sensors. The environmental monitoring was connected to the analyses of water and dissolved gas of the springs performed by ion and gas chromatography method, respectively [58, 59]. Temperature (T in  C) and pH were measured directly on site using a multiparametric PCE-PHD 1 probe. The sampled waters were characterized by temperature ranging between 33.4  C and 34.8  C, pH values between 6.07 and 6.54, relatively high total dissolved solids

1282

S. Bracci et al.

(TDS, up to 2761 mg L 1), and a Ca2+ (HCO3 , SO42 ) composition. Mg2+ and ΣS2 concentrations were relatively high, whereas significant concentrations of Na+, Cl , K+, NO3 , F , and NH4+ were also measured. The composition of the dissolved gas samples was dominated by CO2 (from 4.0 to 6.3 mmol L 1, corresponding to 92–98% of the gaseous phase). A strong relationship between exposition to water and gases and decay phenomena was evidenced. The graphs in Fig. 43.6 show the distribution of the main decay phenomena for each sampled monument. Widespread granular disintegration is due to endogenous heat that, combined with the environmental parameters, determines a strong thermal stress on marble surfaces in the Sanctuary of Apollo, Ploutonion, and Stoa of the Springs. Salt efflorescence, mainly constituted by sulfates (Fig. 43.7d–f), is attributable to the circulation of the thermal spring waters of Hierapolis. The high

Fig. 43.6 Distribution of the main decay phenomena in the studied monuments

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1283

Fig. 43.7 Encrustations of different carbonatic phases (a–b) such as huntite (CaMg3(CO3)4) (c); examples of salt efflorescences (d, e, f); marble and travertine blocks of the Ploutonion completely blackened by crusts of Mn oxides (g, h, i)

concentration of dissolved CO2 in the thermal waters and its chemical nature are responsible of dissolution and corrosion processes and precipitation of various carbonate phases (i.e., calcite, aragonite, huntite) (Fig. 43.7a–c). Actually, biological colonization, although prevalent in the Marble Stoa and Gymnasium, is quite widespread in all the monuments of the archaeological site thanks to the favorable environmental conditions that characterize Hierapolis together with the porosity of the marble used [18]. A strange decay phenomenon was also observed: encrustations, of centimeter thickness, were diffusely present on marble and travertine blocks of the Ploutonion, extracted from wet ground and water in front of the entrance to the grotto (Fig. 43.7g–i). Some of these encrustations were soft and with a subspherical structure in the superficial part of the stone block, while those in the inner areas, in contact with the stone substrate of unaltered marble/travertine, were hard and adherent. Also, the internal walls of the natural cave are entirely blackened and covered by these encrustations. Mn oxides were responsible for the dark color and encrustations [60]. The morphology of the Mn encrustation allows us to identify the role of microorganisms and identify this process as a biomineralization that occurred on both the architectural blocks of the Ploutonion, after their collapse, and the internal walls of the natural cave, whose dark color could have emphasized the suggestion of the holy place. A study of biodiversity of marble surfaces of Gymnasium, North Agora, Marble Stoa, and Central Agora was performed. These monuments are characterized by

1284

S. Bracci et al.

different conservation features, because of both their spatial context (in terms of their position with respect to the seismogenic fault and thermal springs) and the various post-ancient events affecting the buildings. The biodiversity was represented mainly by various types of round-shaped cyanobacteria and green algae, crustose lichens, and one type of moss, except for the Central Agora site, where diatoms, filamentous, and coccoid cyanobacteria were the main constituents [61]. Cyanobacteria and green algae can be associated with a fungus, leading to the formation of lichen. Black patina found in Gymnasium, North Agora, and Marble Stoa is associated with the presence of cyanobacteria and/or dark lichens, which may induce a black hue of the colonized stone due to their development on and inside the stone, penetrating between grains and in fissures, and covering extended areas. Verrucaria sp., a black lichen, and two types of coccoid cyanobacteria were common in all the three sites [61]. A particular example of biological colonization is offered by the remains of the Central Agora that were submerged into a lake formed in the Medieval period, due to the seismic events that greatly modified the hydrological equilibrium of the area [22]. Nowadays, the marble blocks of the ancient porticos are included in the touristic structures of the Pamukkale thermal complex and are one of the main interesting features for the 1.7 million tourists that visit the site every year (Fig. 43.8a) [23]. It represents an important case study concerning the impact of mass tourism on an archaeological site. Such tourism generates problems due to the improper use of the monuments bathing in the lake formed inside the ruins of Roman square. The red patinas (Fig. 43.8b) present on the marble blocks consist of filamentous cyanobacteria (Pseudoanabaena pseudoterminata) and Leptolyngbya sp. spherical unicellular cyanobacteria, and two types of diatoms (Navicula sp. and Surirella sp.) were also observed (Fig. 43.8c–e). All these microorganisms use to develop at the interface between air and water. The variation of interface, due to the presence of numerous visitors, emphasizes this phenomenon. The monitoring of the state of conservation of marble surfaces in different monuments which are at different distances from the seismogenic fault (i.e., St. Philip Church, Ploutonion, Marble Stoa) was performed using noninvasive and nondestructive methodologies such as colorimetric methods and contact sponge method (UNI EN 15886/ 2010; UNI 11432/2011). These parameters indicate the state of conservation of the surfaces and reveal their development over time. In the case of a column in the Marble Stoa, additional cleaning and protection trials were conducted. Two zones in particular were identified, one relative to a lower section excavated in 2005, free from a thick blanket of travertine, and another which was always exposed. The two portions of the column were subdivided into four sections: sub-area cleaned and treated, sub-area cleaned and not treated, sub-area not cleaned and treated, and sub-area not cleaned and not treated. Over the 3 years, all four sub-areas were evaluated for changes in color and water absorption capacity. Both on cleaned and not cleaned sub-areas, in fact, the amount of water absorbed by the treated sub-areas was less than the untreated sub-areas, and this trend has been maintained over time. A diagnostic campaign of state of conservation of mural painting was performed in Insula 104, a residential area dating to the early Byzantine period (5th-6th century

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1285

Fig. 43.8 Tourists bathing in the Central Agora (a); state of conservation of some marble elements (b; red arrows indicate the red patinas at the interface between air and water); stereomicroscopic image of an analyzed sample (c); the biodiversity developed in the glass tubes at the air-water interface (d); the presence of filamentous cyanobacteria such as Pseudoanabaena sp. (e)

AD). On site screening by portable techniques (IR thermography, Quantofix test strips) was performed to better address the microsampling for the laboratory analyses. Sulfates, nitrates, and chlorides are the identified salts, and their presence is mainly due to rising of underground water [62].

43.4.4 Methods for the Study of Ancient Building Materials The identification of the ancient building materials represents an effective tool to reconstruct the economic and social dynamics associated with the creation of the great public complexes that characterize the ancient city. Thanks to the high level of knowledge and conservation of both the monumental complexes and the surrounding territory, Hierapolis represents a perfect case to reconstruct the supply strategies over the centuries. Four main areas of extraction of white and veined marble have been identified in the territory (Fig. 43.1): two close to the city (i) Hierapolis-Gök Dere and (ii) Marmar Tepe (Fig. 43.9a–b), and other two more distant, (iii) Thiounta and (iv) Gölemezli (respectively, 20 km north and 13 km north-west of the city) [63]. An extensive sampling enabled the detailed archaeometric characterization of the extracted building stone by

1286

S. Bracci et al.

Fig. 43.9 Ancient marble quarries on the Marmar Tepe

means of the integration of (i) minero-petrographic, (ii) isotopic, and (iii) cathodoluminescence analyses [64–66]. On this backdrop, the provenance of marbles used in the city was also based on an extensive sampling of architectural artifacts from the public

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1287

monuments and necropolises [67]; in the various monuments, a block of each type (from stylobate to cornice) was sampled (Fig. 43.10a, c–g). In the well-preserved and deeply studied monuments, the sampling was even more abundant, and in the case of the scaenae frons of the Theatre, all the blocks of the first order were analyzed. The archaeometric analyses offered interesting information on the building stone procurement strategies, especially between the first Imperial age and the Severan era [68] (Fig. 43.10b). In particular, 91.5% of the marbles used in the city and necropolises resulted as extracted in the local quarries of Marmar Tepe, Hierapolis-Gök Dere, and Thiounta. In particular, Marmar Tepe and Hierapolis-Gök Dere, the two extraction areas closest to the city, provided 73% of the material, while Thiounta provided 17% of the marble used. An extremely limited quantity of marble (around 2%) came from the Gölemezli and the Denizli quarries. The material imported from the more distant quarries of Dokimeion and Aphrodisias amount to 7% overall. In addition to the historical conclusions that have been drawn by the archaeometric analysis, it should be pointed out the importance of the methodological approach applied to the sampling and archaeometric analyses. In fact, the project was based on a strongly contextual approach, which means (i) precise georeferencing and documentation of the sampling points in the quarries and the archaeological area, (ii) reconstruction of the function and position of the sampled and studied artifacts within the architectural layout of the monuments, and (iii) paying close attention to the contexts of primary and secondary use of the materials. All this required the integration of the spatial component in the documentation and the management of all the entities described in the project’s records (monuments, artifacts, and samples) by means of a specific online geodatabase (http://antares5.ibam.cnr.it) [69, 70]. This digital archive was based on an open-source database server and was structured in three interconnected records: (i) “context,” concerning the contexts investigated, i.e., the marble quarries and the urban and funerary monuments of Hierapolis; (ii) “object,” concerning the artifacts within the contexts of investigation that were subject to sampling; and (iii) “sample,” concerning the results of the archaeometric analyses conducted on the individual samples, subdivided into two sub-records, one for the analyses designed to characterize the quarries and determine the provenance of the archaeological artifacts of Hierapolis, and one for the analyses performed to determine their state of conservation (Fig. 43.11). A webGIS interface makes it possible to consult the archaeological and archaeometric data by means of geospatial queries, starting from the position of the data in space.

43.4.5 Methods for the Reconstruction of the Ancient City Landscape The reconstruction of the city’s ancient landscape represents the ultimate goal of our archaeological research. The investigations on the geological setting of Hierapolis and the study of its natural resources were integrated within the unique framework of the digital archaeological cartography of the city and the necropolises [5]. However, the 2D representation of the city plan is not enough to offer an exhaustive image of the

1288

S. Bracci et al.

Fig. 43.10 Sampled artifacts from the North Agora (a, c–g) and percentages of marble used in Hierapolis between the Hellenistic age and the Byzantine period (b)

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1289

Fig. 43.11 Marmora Phrygiae geodatabase: Sample Table – State of Conservation, results of a query and images connected to this record

1290

S. Bracci et al.

complex experience of the ancient urban landscape. For this reason, the further step is represented by the architectural study of the individual monumental complexes. In this framework, the reconstruction of their elevation was conducted by means of surveys (both manual and instrumental) and the careful study of the conserved architectural elements [71]. The instrumental surveys (GPS, Total Station) were integrated with 360-degree scans and image-based techniques such as 3D photogrammetry (Camera Scanner), producing geometric models, complete with texture. After the survey, a painstaking examination of the architectural blocks was performed in order to define their mutual relationships in the vertical and horizontal rows, according to their technical features, their decorative patterns, and the position of the collapsed materials. In this framework, great support was also offered by the Virtual Archaeology [72]. The workflow foresaw the creation of a reasonable number of digital models of individual monuments. With respect to the 2D reconstructions, texturing and lighting of the 3D objects allowed us to achieve the hyper-realistic rendering of the original colors and surfaces of the monuments. In order to obtain a rendering more faithful to the original, specific textures were created for every monument by means of highresolution photographic sampling of the surfaces, thus reproducing the rich polychromy of the decorative apparatuses of the monuments. To reconstruct the image of the entire city, the individual models were inserted within a user experience platform. It allows us to walk through a hyper-realistic virtual architectural space, which visualizes the spaces enclosed between the volumes, the paths, the colors, and the light-shadow features. The chosen representation system was based on Quick Time Virtual Reality technology, that is, a virtual tour of the city using cylindrical or spherical panoramic images according to the well-known Google Street View method (Fig. 43.12) [73]. The information and visualizations of monuments contained in the virtual tour are inserted into a dynamic and interactive context which provides a better understanding of the urban appearance of the city as a whole, placing the buildings in the surrounding environment and reconstructing the perception of spaces and volumes that have been defaced or are no longer visible. At the same time, it allows us to comprehend the topographic evolutions and transformations of Hierapolis over the centuries, thanks to the reconstruction of the different building phases, accessible by means of switches included within the panoramas. In addition, the platform represents a synthesis tool in which three-dimensional models of monuments also integrate informative content: in fact, photos of the actual statue of the buildings and short descriptions of their history were offered to the user, who is able to access these contents via linkable hot spots.

43.5

Conclusions

The experience of the Marmora Phrygiae project showed the potentiality of archaeological sites characterized by a high level of knowledge and conservation of both the monumental complexes and their surrounding territory. The synergy between humanistic scholars and scientific researchers meant that the archaeometric or geophysical investigations were designed to resolve specific archaeological

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1291

Fig. 43.12 Graphical user interface of the Hierapolis virtual tour: Frontinus Gate (a), Temple A in the Sanctuary of Apollo (b), Tholos in the Ploutonion (c)

questions and issues of the historical reconstruction. In addition, the archaeometric results were enriched in many cases with a diachronic perspective, which is evident in both the cited studies of the procurement of the marbles and the exploitation of the

1292

S. Bracci et al.

quarries, but also in the reconstruction of the geomorphological and tectonic setting, thanks to the stratigraphic information from archaeological excavation. An important result of the project’s multidisciplinary approach is the abundant data acquired regarding the state of conservation of the marbles of Hierapolis and the types of threat to which they are subject. On various scales, the research studied the effects of the thermal spring waters and gaseous emissions, as well as the forms of decay and human hazards threating the ancient buildings and the territory surrounding the city. The knowledge acquired in these fields of investigation provided the foundations on which to plan the future conservation of the monuments and the site management. Acknowledgments The authors would like to thank Francesco D’Andria, former Director of the Italian Archaeological Mission in Hierapolis in Phrygia, for having supported the Marmora Phrygiae project, as well as for providing a wealth of advice, suggestions, and ideas for research. In addition, we express our thanks to the Direction of the Archaeological Museum in Hierapolis and to the representatives of the Ministry of Culture and Tourism of the Republic of Turkey, for the generosity and enthusiasm with which they facilitated the research.

References 1. D’Andria F (2003) Hierapolis di Frigia (Pamukkale). Guida archeologica. Ege Yayınları, Istanbul 2. Ritti T (2017) Storia e istituzioni di Hierapolis, Hierapolis di Frigia IX. Ege Yayınları, Istanbul 3. D’Andria F (2013) Il Ploutonion a Hierapolis di Frigia. Istanbuler Mitteilungen 63:157–217 4. D’Andria F (2018) The Ploutonion of Hierapolis in light of recent research (2013–2017). J Roman Archaeol 31:91–129 5. Scardozzi G (ed) (2015) Nuovo Atlante di Hierapolis di Frigia. Cartografia archeologica della città e delle necropoli, Hierapolis di Frigia VII. Ege Yayınları, Istanbul 6. D’Andria F, Caggia MP (eds) (2007) Hierapolis di Frigia I. Le attività delle campagne di scavo e restauro 2000–2003. Ege Yayınları, Istanbul 7. D’Andria F, Caggia MP, Ismaelli T (eds) (2012) Hierapolis di Frigia V. Le attività delle campagne di scavo e restauro 2004-2006. Ege Yayınları, Istanbul 8. D’Andria F, Caggia MP, Ismaelli T (eds) (2016) Hierapolis di Frigia VIII. Le attività delle campagne di scavo e restauro 2007–2011. Ege Yayınları, Istanbul 9. Ismaelli T, Scardozzi G (2016) The North Agora: the building site and the provenance of marbles. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 277–286 10. Scardozzi G (2012) New data on the North Theatre of Hierapolis in Phrygia: archaeological, topographical and geophysical surveys. In: Masino F, Mighetto P, Sobrà G (eds) Restoration and management of ancient theatres in Turkey. Methods, research, results. Congedo Editore, Galatina, pp 217–235 11. Ismaelli T, Scardozzi G, Sobrà G (2016) New data on the building site of the Theatre of Hierapolis. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 305–328 12. Ismaelli T (2017) Il Tempio A nel Santuario di Apollo. Architettura, decorazione e contesto, Hierapolis di Frigia X. Ege Yayınları, Istanbul 13. Campagna L (2018) Il Ninfeo dei Tritoni, Hierapolis di Frigia XI. Ege Yayınları, Istanbul

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1293

14. D’Andria F (2016–2017) “Hierapolis alma Philippum”. Nuovi scavi, ricerche e restauri nel Santuario dell’Apostolo, Atti della Pontificia Accademia Romana di Archeologia. Rendiconti 89:129–177 15. Marabini S (2016) Geological and geomorphological setting of the area surrounding Hierapolis. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 79–84 16. Kumsar H, Aydan Ö, Şimşek C, D’Andria F (2016) Historical earthquakes that damaged Hierapolis and Laodikeia antique cities and their implications for earthquake potential of Denizli basin in western Turkey. Bull Eng Geol Environ 75(2):519–536 17. Andreotti A, Baroni S, Bonaduce I, Bozza S, Cantisani E, Ismaelli T, Vettori S (2018) Ancient restorations at Hierapolis of Phrygia (Denizli, Turkey): interdisciplinary research on materials and technologies. J Archaeol Sci Rep 21:862–871 18. Vettori S, Cabassi J, Cantisani E, Riminesi C (2019) Environmental impact assessment on the stone decay in the archaeological site of Hierapolis (Denizli, Turkey). Sci Total Environ 650: 2962–2973 19. Pfanz H, Yüce G, Gülbay AH, Gökgöz A (2019) Deadly CO2 gases in the Plutonium of Hierapolis (Denizli, Turkey). Archaeol Anthropol Sci 11(4):1359–1371 20. Kele S, Özkul M, Gökgöz A, Fórizs I, Baykara MO, Alçiçek MC, Németh T (2011) Stable isotope geochemical and facies study of Pamukkale travertines: new evidences of low-temperature non-equilibrium calcite-water fractionation. Sediment Geol 238:191–212 21. Marabini S, Scardozzi G (2015) La ricerca geo-archeologica a Hierapolis. In: Scardozzi G (ed) Nuovo Atlante di Hierapolis di Frigia. Cartografia archeologica della città e delle necropoli, Hierapolis di Frigia VII. Ege Yayınları, Istanbul, pp 227–268 22. Ismaelli T, Scardozzi G, Bozza S (2017) L’Agorà Centrale di Hierapolis di Frigia. Mare Internum 9:113–141 23. D’Andria F (2018) Hierapolis in Phrygia. Tourism-related use and abuse of an archaeological site. In: Piesker K, Akan B, Göçmen D, Altay ST (eds) Heritage in context 2, archaeology and tourism. Ege Yayınları, Istanbul, pp 201–230 24. Barone PM, Scardozzi G (2016) Optical high-resolution satellite imagery for the study of the ancient quarries at Hierapolis. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 657–668 25. Scardozzi G (2007) Remote sensing and archaeological survey in the Hierapolis of Phrygia territory, Turkey. In: Remote sensing for environmental monitoring, GIS applications, and geology. Conference 6749 “remote sensing and archaeology” (SPIE Europe Remote Sensing 2007) 6749/04. SPIE, pp 1–12 26. Scardozzi G (2012) Integrated methodologies and technologies for the reconstruction of the ancient topography of Hierapolis in Phrygia and its surrounding territory (Turkey). In: Proceedings of the 5th international congress “science and technology for the safeguard of cultural heritage in the Mediterranean Basin”. AIC, Roma, pp 192–202 27. Lasaponara R, Masini N, Scardozzi G (2008) New perspectives for satellite-based archaeological research in the ancient territory of Hierapolis (Turkey). Adv Geosci 19:87–96 28. Lasaponara R, Masini N, Scardozzi G (2010) Elaborazione di immagini satellitari ad alta risoluzione e ricognizione archeologica per la conoscenza degli insediamenti rurali del territorio di Hierapolis di Frigia (Turchia). In: D’Andria F, Malfitana D, Masini N, Scardozzi G (eds) Il dialogo dei saperi. Metodologie integrate per i Beni Culturali. Edizioni Scientifiche Italiane, Napoli/Roma, pp 479–494 29. Lasaponara R, Leucci G, Masini N, Persico R, Scardozzi G (2016) Towards an operative use of remote sensing for exploring the past using satellite data. The case study of Hierapolis (Turkey). Remote Sens Environ 174:148–164 30. Di Giacomo G, Scardozzi G (2015) Riprese aeree con pallone aerostatico frenato e droni. Esperienze tra ricerca e documentazione archeologica a Hierapolis di Frigia. Archeologia Aerea 9:101–116

1294

S. Bracci et al.

31. Castrianni L, Di Giacomo G, Ditaranto I, Scardozzi G (2010) High resolution satellite orthoimages for archaeological research: different methods and experiences in the near and Middle East. Adv Geosci 24:97–110 32. Di Giacomo G, Ditaranto I, Scardozzi G (2011) Cartography of archaeological surveys taken from an Ikonos stereo pair: a case from Hierapolis of Phrygia territory (Turkey). J Archaeol Sci 38:2051–2060 33. Scardozzi G (2008) Old high-resolution satellite images for landscape archaeology: case studies from Turkey and Iraq. In: Remote sensing for environmental monitoring, GIS applications, and geology. Conference 7110 “remote sensing and archaeology” (SPIE Europe Remote Sensing 2008) 7110/03. SPIE, pp 1–14 34. Lasaponara R, Masini N, Scardozzi G (forthcoming) Potenzialità dei dati satellitari SAR Cosmo-SkyMed per l’archeologia: primi risultati a Hierapolis di Frigia (Turchia). Archeol Aerea 10:59–68 35. Scardozzi G (2007) Hierapolis di Frigia. Applicazioni informatiche alle ricognizioni archeologiche e telerilevamento da satellite: l’esempio degli acquedotti della città. Archeologia e calcolatori 18:331–353 36. Scardozzi G (2010) Hierapolis di Frigia, dalle cave ai cantieri di demolizione. L’approvvigionamento di materiali lapidei nella città di età imperiale e proto-bizantina. In: Camporeale S, Dessales H, Pizzo A (eds) Arqueología de la construcción II. Los procesos constructivos en el mundo romano: Italia y provincias orientales. Istituto de Arqueología de Mérida, Madrid/ Mérida, pp 351–374 37. Scardozzi G (2012) Ancient marble and alabaster quarries near Hierapolis in Phrygia (Turkey): new data from archaeological surveys. In: Garcia MAG, Lapuente P, Rodà I (eds) Interdiscplinary studies on Ancient Stone. Proceedings of the IX ASMOSIA international conference. Institut Català d’Aqueologia Clàssica, Tarragona, pp 573–583 38. Scardozzi G (2012) Un nuovo miliario dal territorio di Hierapolis di Frigia. Contributo alla ricostruzione della viabilità antica nella valle del Lykos. In: D’Andria F, Caggia MP, Ismaelli T (eds) Hierapolis di Frigia V. Le attività delle campagne di scavo e restauro 2004–2006. Ege Yayınları, Istanbul, pp 739–766 39. Scardozzi G (2012) Ricognizioni archeologiche nel territorio di Hierapolis (campagne 2005–2006): gli acquedotti, le cave di materiali lapidei, gli insediamenti rurali, i tumuli funerari. In: D’Andria F, Caggia MP, Ismaelli T (eds) Hierapolis di Frigia V. Le attività delle campagne di scavo e restauro 2004–2006. Ege Yayınları, Istanbul, pp 109–143 40. Scardozzi G (2013) Nuovi dati sulla topografia antica di Hierapolis di Frigia. Atlante Tematico di Topografia Antica 23:25–59 41. Scardozzi G (2014) Topografia antica del territorio di Hierapolis di Frigia: ricognizioni archeologiche negli altopiani di Uzunpınar e Çal. Scienze dell’Antichità 20:95–122 42. Scardozzi G (2010) Oil and wine production in Hierapolis of Phrygia and its territory during Roman and Byzantine age: documentation from archaeological excavations and surveys. In: Aydinoğlu Ü, Şenol AK (eds) Olive oil and wine production in Anatolia during antiquity. Symposium proceedings. Kilikia Arkeolojisini Araştırma Merkezi, Istanbul, pp 277–302 43. Scardozzi G (2016) Necropoleis from the territory of Hierapolis in Phrygia: new data from archaeological surveys. In: Brandt JR, Hagelberg E, Bjørnstad G, Ahrens S (eds) Life and death in Asia Minor in Hellenistic, Roman, and Byzantine times: studies in archaeology and bioarchaeology. Oxbow Books, Oxford/Philadelphia, pp 19–38 44. Scardozzi G (2013) I santuari del territorio di Hierapolis di Frigia: nuovi dati dalle ricognizioni archeologiche. In: Giardino L, Tagliamonte G (eds) Archeologia dei luoghi delle pratiche di culto. Atti del convegno. Edipuglia, Bari, pp 69–88 45. Scardozzi G (2013) Antiche divisioni agrarie in Asia Minore: problematiche e strumenti di ricerca. Archeol Aerea 7:124–148 46. Scardozzi G (2007) Ricerche topografiche e telerilevamento. In: D’Andria F, Caggia MP (eds) Hierapolis di Frigia I. Le attività delle campagne di scavo e restauro 2000–2003. Ege Yayınları, Istanbul, pp 67–86

43

Multidisciplinary Approaches to Study Ancient Cities in a Seismic Region

1295

47. Ditaranto I (2016) Ancient marble quarries in the territory of Hierapolis. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 87–100 48. Scardozzi G (2016) The alabaster quarries of Hierapolis. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 141–165 49. Cantisani E, Scardozzi G (2016) The polychromatic breccia of Hierapolis: quarries, use in construction sites, and alteration phenomena. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in southwestern Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 167–180 50. Marabini S (2015) Note illustrative della carta geologica di Hierapolis. In: Scardozzi G (ed) Nuovo Atlante di Hierapolis di Frigia. Cartografia archeologica della città e delle necropoli, Hierapolis di Frigia VII. Ege Yayınları, Istanbul, pp 7–11 51. Di Giacomo G, Ditaranto I, Miccoli I, Scardozzi G (2015) La faglia sismica di Hierapolis. In: Scardozzi G (ed) Nuovo Atlante di Hierapolis di Frigia. Cartografia archeologica della città e delle necropoli, Hierapolis di Frigia VII. Ege Yayınları, Istanbul, pp 207–226 52. Leucci G (2015) Geofisica applicata all’archeologia e ai beni monumentali. Flaccovio Dario, Palermo 53. Semeraro G (2007) Ricerche archeologiche nel Santuario di Apollo (Regio VII) 2001–2003. In: D’Andria F, Caggia MP (eds) Hierapolis di Frigia I. Le attività delle campagne di scavo e restauro 2000–2003. Ege Yayınları, Istanbul, pp 169–209 54. Semeraro G (2012) Ricerche nel Santuario di Apollo. In: D’Andria F, Caggia MP, Ismaelli T (eds) Hierapolis di Frigia V. Le attività delle campagne di scavo e restauro 2004–2006. Ege Yayınları, Istanbul, pp 293–324 55. Leucci G, Ismaelli T, Scardozzi G (2016) Geophisical surveys at Hierapolis. New data on monuments and tectonic context. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in southwestern Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 635–656 56. Ismaelli T (2009) Il monopteros del Santuario di Apollo a Hierapolis di Frigia. Ricerche sull’oracolo alfabetico. Istanbuler Mitteilungen 59:131–192 57. Di Giacomo G, Vettori S, Cantisani E, Riminesi C, Scardozzi G (2016) Documentation and analysis of the seismic fissures of the Stoa of the Springs at Hierapolis using thermographic imaging from a tethered aerostatic balloon. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in southwestern Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 625–634 58. Montegrossi G, Tassi F, Vaselli O, Bidini E, Minissale A (2006) A new, rapid and reliable method for the determination of reduced Sulphur (ΣS2-) species in natural discharges. Appl Geochem 21:849–857 59. Vaselli O, Tassi F, Montegrossi G, Capaccioni B, Giannini L (2006) Sampling and analysis of volcanic gases. Acta Vulcanol 18(1-2):65–76 60. Vettori S, Cantisani E, Chelazzi L, Cuzman OA, Gatta GD, D’Andria F (2019) The dark colour of the Ploutonion at Hierapolis of Phrygia (Turkey). Archaeometry 61(2):296–308 61. Cuzman OA, Leal AS, Schaffar S, Vettori S, Cantisani E (2016) Notes on the stone biodiversity of Hierapolis monuments. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 609–618 62. Vettori S, Bracci S, Cantisani E, Riminesi C, Sacchi B, D’Andria F (2016) A multi-analytical approach to investigate the state of conservation of the wall paintings of Insula 104 in Hierapolis (Turkey). Microchem J 128:279–287 63. Scardozzi G (2016) General considerations on the ancient marble quarries in the territory of Hierapolis and in the southern sector of the Denizli basin: between topography and

1296

S. Bracci et al.

archaeometry. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 131–140 64. Brilli M, Conte AM, Giustini F, Mercadal MPL, Melica D, Quarta G, Plumed HR, Scardozzi G (2016) Archaeometric characterization of the white marbles from the ancient quarries in the territory of Hierapolis and in the southern sector of the Denizli basin, with an appendix on the Aphrodisian marble. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 101–118 65. Koralay T (2016) Archaeometric characterization of the recently discovered Yenişehir marble quarry in the Denizli region. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 119–130 66. Long LE (2012) The regional marble quarries. In: Ratté C, De Staebler P (eds) The Aphrodisias regional survey, Aphrodisias V. Philipp von Zabern, Darmstadt, pp 165–201 67. Quarta G, Brilli M, Giustini F, Melica D, Lapuente Mercadal MP, Royo Plumed H, Ismaelli T, Scardozzi G (2016) Provenance of the marbles used in the monuments of Hierapolis through an archaeometric approach: petrography, isotopes and cathodoluminescence. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 183–230 68. Scardozzi G (2016) Marble supply strategies in the building sites of Hierapolis. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 421–435 69. Di Giacomo G (2016) The on-line platform of the Marmora Phrygiae Project. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 33–41 70. Scardozzi G, Di Giacomo G, Ismaelli T (eds) (2018) Marmora Phrygiae. Il geodatabase delle cave di marmo e dei cantieri costruttivi di Hierapolis di Frigia: archeologia, archeometria e conservazione. Edizioni CNR-IBAM, Roma 71. Ismaelli T (2016) Architecture and decoration in context: a cross analysis of the ancient building site in five steps. The case study of the Severan temple a in the Sanctuary of Apollo. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 339–364 72. Limoncelli M (2016) The contribution of virtual archaeology to the Marmora Phrygiae Project. In: Ismaelli T, Scardozzi G (eds) Ancient quarries and building sites in Asia Minor. Research on Hierapolis in Phrygia and other cities in south-western Anatolia: archaeology, archaeometry, conservation. Edipuglia, Bari, pp 43–60 73. Limoncelli M (forthcoming) Virtual Hierapolis, Hierapolis di Frigia XIII. Ege Yayınları, Istanbul

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

44

Laura Brambilla

Contents 44.1 44.2 44.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Documentation and Characterization of a Can’s Collection . . . . . . . . . . . . . . . . . . . . . Results of the Survey of Cans in Museums . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.3.1 Cans Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.4 Tinplate Materials Identification and Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.5 Corrosion of Tinplate in Complexing Acid Media Mimicking Real Cans’ Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1298 1299 1303 1304 1313 1320 1328

Abstract

This chapter is dedicated to the development of a methodology for the study of the degradation of complex composite artifacts. Among those, food cans, made of tinplate and still retaining their original content, are a case study representative of this category of objects. The research was divided in four phases: (1) the development of a method for documenting the state of conservation of food cans in collections; (2) the condition report, carried out on around 150 cans belonging to five different Swiss collections, and aiming at assessing the main degradation occurring on these artifacts; (3) the characterization, using electrochemical techniques and surface analyses, of cans’ constituting materials on commercially available samples as mock-ups; and (4) the investigation of corrosion mechanisms of tinplate in complexing acid media, used as environments mimicking the real cans’ contents, i.e., food. This study, conducted in the framework of the Swiss project titled CANS (Conservation of cAns in collectioNS), was the first interdisciplinary and comprehensive research on the long-term interaction between the tinplate and the food. L. Brambilla (*) Haute Ecole Arc Conservation-restauration, HES-SO University of Applied Sciences and Arts Western Switzerland, Neuchâtel, Switzerland e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_44

1297

1298

44.1

L. Brambilla

Introduction

Nowadays, museums conserve not only what is commonly accounted as artworks, such as sculptures, paintings, or archaeological objects, but also every day’s life artifacts, considered as witnesses of our society and its changes. Among these objects, some of the most unusual are food cans, still filled with their original contents. A first systematic study of food cans in collections, their state of conservation, and degradation phenomena was carried out in the framework of the CANS project (Conservation of cAns in collectioNS), granted by the Swiss National Science Foundation between 2014 and 2017 [1]. A survey conducted by Olivier Schinz, ethnologist at MEN (Musée d’Ethnographie de Neuchâtel, Switzerland) during the CANS project, revealed that in museums around the world are conserved and exposed more than 20,000 cans. Amid those, around 6000 still retain their original content [2]. These cans date from the mid of the nineteenth century to nowadays, as many museums are still collecting cans. There are different motivations for the conservation of food cans in museums collections, among them: – Cans represent a specific local industrial production, as it is the case of vegetables canning worldwide or fish industry along the coasts of many European countries. – They may be kept as a proof of production or production change of a precise brand. Most of the main canned food producers have their own collection. – They may be witnesses of food habits during wars or expeditions, including Earth’s poles explorations and conquest of space. – They may carry a symbolic meaning as demonstrated by the abundance of artworks including cans, created by several artists of contemporary art (Manzoni, Warhol, Ben, Spoerri, and Pope. L, just to cite some of them). A food can is a complex composite object, from a material point of view, as shown in Fig. 44.1, consisting of: • The metallic container that is, itself, a composite material made usually of a mild steel base sheet covered on both sides with tin. A polymeric coating is often applied on top of the tin on the internal part of the can to isolate the metal from the food. • A large variety of food contents, each presenting its own physico-chemical specificities. • An external label that can be directly printed or painted on the container or, more commonly, made of paper. The conservation of composite objects, such as cans, is a real challenge for museums conservators. The preventive conservation measures to be taken, in

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1299

Fig. 44.1 Can as a composite object and its composing parts. In the green box, it is emphasized the double seam

terms of environmental control (T, RH%, light), as well as the conservation treatments, are indeed different for the diverse materials composing the artifact. For collections of cans still filled with their original content, the conservation is even more problematic because corrosion phenomena, often severe for the artifacts, occur due to interactions on the long term between the organic content and the metallic container. In addition, environment plays an important role. Therefore, the development of a proper methodology for the characterization and the monitoring of the cans of a collection and their materials is of utmost importance in order to follow and anticipate their changes during time. This approach will be described in the following pages.

44.2

Documentation and Characterization of a Can’s Collection

The condition report is the first evaluation step commonly carried out in the field of cultural heritage. It aims at determining the state of conservation of a collection. This procedure allows to describe the conditions of an artifact in term of stability and state of conservation at a certain moment in time [3]. Since no protocol was available, to the knowledge of the CANS project participants, for an appropriate condition report

1300

L. Brambilla

regarding heritage cans, a specific procedure was developed for these objects, in order to observe and measure the most relevant characteristics [4] attesting of the artifact condition. The condition report should be carried out using simple instruments (photographic equipment, portable microscopy, caliber, external caliber) that may be coupled with portable analytical techniques, when available, such as portable X-Ray Fluorescence Analysis (p-XRF), if analyses on the composition of the container’s materials are necessary. The developed protocol includes the following steps: I. Record the can’s brand, weight, content, and year of production or expiration date. • The data collected during this first phase should be registered in a spreadsheet that can afterward be included in the museum’s archives. The relevant information to be reported in the condition report spreadsheet for a collection of cans are reported in Table 44.1. • The date or period of production is of utmost importance to understand the manufacturing technology used. For example, it may help for discriminating between hot dipping or electroplating tinning, where: Table 44.1 Criteria used to build the condition report spreadsheet Inventory of cans specifications

Information reported on the label

Measurements of the can

Description of the alterations observed on the can

Alterations of the metal

Alterations of the label

Alterations of the content Presence of deposits on the external surface

Brand Content Year of production Expiring date Weight of the content Price Weight Height Diameter (both top and bottom, if different) Height at the center Corrosion Perforation Swelling Mechanical damages Stains Detachment Embrittlement Rips Holes Loss of readability Complete loss Leakage Dehydration Dust Dried content due to leakage External deposits of unknown nature

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1301

– The hot-dipping tinning, which was used to produce the first industrial cans, corresponds to a thicker layer of tin, similar on the two sides of the can (internal, in contact with the food, and external). – The electroplating procedure itself is more recent and corresponds to a thinner layer of tin, often applied in different thicknesses internally and externally, with the internal layer usually thicker than the external one. • The content of the can may give important information about the eventual presence of specific oxidizers or detinners in the food that can potentially attack the tinplate and promote or accelerate corrosion. This is typically the case, for example, for certain green vegetables, like green beans or spinach, containing high amounts of nitrates that promote tin dissolution [5]. More detailed information about the food corrosivity is reported below in Sect. 44.5. • The weight declared on the label of the can may be compared to the weight measured during the condition report. II. Carry out comprehensive photographic documentation of the can from all its sides. In order to document the visual aspect of the can in its current state, a complete photographic documentation of its entire surface should be carried out. This documentation that can also be included in the inventory files of the museums allows the comparison of the evolution of the state of conservation of the can at different moments of time. For example, the appearance or enlargement of a stain on the paper label or of a corrosion spot may correspond to the evolution of degradation phenomena, such as perforation of the container. The illumination condition is a critical parameter in order to be able to compare pictures taken at different times. For the light exposure to be reproducible, the use of a foldable light box with dark or white background is a prerequisite, as illustrated in Fig. 44.2. In addition, a light box provides a diffuse light and reduces parasitic reflections from the metal surface of the can. Photography can be combined with portable microscopy, so as to document details on the external surface of the can, such as corrosion spots, or potential perforations of the container. For this purpose, the USB microscopes available on the market are a convenient choice. Images taken with a USB microscope on cans, emphasizing the presence of corrosion are presented in Fig. 44.3. III. Measure of can’s dimensions: weight, height, diameter, and height at the center of the can. These data, in addition to provide statistical outcomes on collections, may be used for monitoring the evolution of the can’s condition. This is especially the case of the weight of the can and its height at the center (i.e., on the axis of revolution of the cylindrical shape). – The weight of the whole can (container and content) is measured using standard methods. It should be compared to the weight of the content declared on the label in order to detect possible important leakages.

1302

L. Brambilla

Fig. 44.2 Setup for photographic documentation of cans using a light box with white background

Fig. 44.3 Images taken on cans using a USB microscope, allowing to confirm the presence of corrosion spots (on the left) or a better characterization of a corroded area (on the right)

– The height at the center of the can is measured using an outside caliper, like the one shown in Fig. 44.4. The development of gas, consequence of internal corrosion, may be detected by measuring swelling. The measurement by means of an outside caliper allows one to detect the presence of swelling before it is perceivable by human eye. The periodic monitoring of these two parameters is particularly important. The developed protocol suggests to perform these measurements every 6 months in order to detect swelling or leakage at a very early stage. IV. Perform observation and detailed description. A detailed description of the external aspect of all the surface of the can should be reported in the spreadsheet. This step of the condition report can

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1303

Fig. 44.4 Measurement of the height of the can at the center, using an outside caliper (Museum Burghalde, Lenzburg)

arduously be standardized and is left to the sensitivity and expertise of the operator. However, this step is of utmost importance in order to leave a description of the condition of the cans at a certain moment in time. A specific terminology for the description of the alteration has to be developed, both for the label and for the metallic container. This description can be supported by microscopic images of the surface of the can.

44.3

Results of the Survey of Cans in Museums

Following the method described in the previous section, a survey has been conducted on around 150 cans belonging to 5 Swiss museums [4, 6]. The aim of this survey was to detect and describe the main problems of full cans in collections, as degradations of this kind of artifacts were not systematically described before. The museums were selected for the number and variety of full cans present in their collections. As previously stated, even if many different kinds of museums conserve cans, most of them prefer to remove the content for the sake of safety and conservation. The Swiss museums conserving cans with their original content are the following: – The Alimentarium in Vevey (Canton of Vaud) is the first museum in the world exclusively dedicated to food and nutrition, open since 1985. The museum is managed by the Nestlé Foundation and presents permanent and temporary exhibitions about contemporary and ancient eating habits from all around the world. Cans, either full or empty, are part of both permanent and temporary exhibitions, even if most of the full cans are conserved in the storage facilities due to their unstable conditions. Around 55 cans, dating from the 1890s to nowadays, were analyzed in this collection.

1304

L. Brambilla

– The Museum Burghalde in Lenzburg (Canton of Aargau) is a small local museum. In 2010 the Hero Company, an important Swiss brand, founded in Lenzburg in 1886 and producing food goods donated to the museum its archives. Part of the Hero archives was exhibited from 2011 to 2013 at the Museum Burghalde, during the exhibition entitled “Hero, seit 1886 in aller Munde” (Hero, from 1886 on everyone’s lips). Hero archives included cans with and without their original content. Around 65 cans, dating from the 1940’s to nowadays, except one produced in 1886 (first day of cans’ production by Hero in Lenzburg), were analyzed in this collection. – The Ortsmuseum in Küsnacht (Canton of Zurich) is another small local museum. In this museum, a historical store (“Tante Emma-Laden”) has been recreated, displaying products commonly sold in the 1950s. These include some cans of concentrated milk. Around 15 cans were analyzed in this collection dating probably from the 1970s (almost no information is available concerning these artifacts). – The HAM foundation in Thun (Canton of Bern) is the museum of the Swiss Army. Within military equipment there is also food ration, including cans, even if their number is limited. Around five cans were analyzed in this collection dating probably from the 1990s. – The Historical Museum of Bern (Canton of Bern) is a museum dedicated to Swiss and worldwide history. In the collection of this museum, there is a batch of food and food packaging, including cans, dating from the 1990s. There is no trace of precise information regarding the original exhibition for which these objects were acquired by the museum. Around 10 cans were analyzed in this collection. It has to be mentioned that not only ancient cans are present in collections. Museums like the Alimentarium still collect cans; therefore it is also possible to find recently produced cans in collections.

44.3.1 Cans Degradation The principal degradation processes faced by cans in collections concern the three main parts composing the cans itself: – The content – The label – The container On the one hand, the variety of different contents available on the market is so broad that it is not possible to assess the degradation of all of them. Here below only some hints on the possible degradations of the content will be provided. The degradation of the paper is largely described in specialized literature [7], and therefore only cans-specific observations are reported here. On the other hand, the degradation of the container is the most destructive process, for the can considered is

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1305

a single composite object. Therefore, the analytical characterization of the container’s materials and of the degradation processes will be debated in a more comprehensive manner.

Degradation of the Content The degradation of the content occurs due to the aging of food. It is highly probable that a separation of different phases may occur for certain kind of food, even if documenting this phenomenon would require techniques that are hardly accessible by museums, as the food remains hidden inside the metallic container. For example, this is the case of condensed milk. The content of a can of condensed milk was analyzed some years ago by the Swiss National Museum by FTIR spectroscopy. The resulting spectra showed explicitly that two phases were present inside the can, one more rich in fatty acids and sugar and another one more rich in water. This separation of phases, and therefore the formation of an interface, is considered as the possible causes of localized corrosion formation. A degradation of the content that is more common to observe in collections is the dehydration, partial or total, of the food. This can occur only if there is an exchange between the inside of the can and the external atmosphere, i.e., most commonly, a perforation. When the metallic container is perforated, the content starts to pour out, potentially damaging the label if the leakage occurs from the cylinder part (see following paragraph). In some cases, depending on the type of content, only the liquid, water-based, part of the content pours out or evaporates, while the solid part remains inside the can (it is the case, e.g., of fruit in syrup or vegetables in water). As mentioned in Sect. 44.2, the presence of leakage can be detected by weighting periodically the cans and comparing their weight with the one, announced on the label, of the content. Degradation of the Label The main degradations observed on cans’ labels are spots and stains. They can result from the contact with an external agent (water, humidity, dust, rust, food), like in Fig. 44.5, or due to the leakage of the content following the perforation or bursting of the container. In the case of container perforation, the contact between the paper and the food content may lead to the formation of very large and dark stains, like in Fig. 44.6, and to the possible development of mold. The risk of complete loss of the label in this case is really high, due to the embrittlement of the paper following the contact with food and the consequent breakage of the label, as illustrated in Fig. 44.7. An additional degradation phenomenon, i.e., aging and embrittlement of the paper, is also often observed due to non-adapted conservation conditions. This aging can lead to the yellowing of the paper and the formation of tears and rips, like in Fig. 44.5. The paper embrittlement is really common as labels were produced with middle
/low-quality paper not destined to last in time longer than the shelf life of the food can. However, the label represents the most important part of the can, from the museums point of view, as it is the one carrying the information (content, brand).

1306

L. Brambilla

Fig. 44.5 Can presenting yellowing of the paper and stains due to external factors, such as humidity or dust (Alimentarium, Musée de l’alimentation, une fondation Nestlé, Vevey)

Also from an aesthetic point of view, the label has a high cultural value, as it is almost the only way to recognize a can from another. As previously mentioned, labels are most of the time, but not always, made of paper. They can also be printed directly on the metal, with serigraphy or lithography techniques, or painted. Some cans with painted labels are present in Swiss collections. The painting present specific degradation problems like weakening, flaking, and detachment due to the same factors causing paper embrittlement, i.e., environmental conditions, underlying corrosion, or contact with corrosive agents including food, like in Fig. 44.8.

Degradation of the Container The primary degradation of the metallic container results from corrosion. Corrosion can affect both internal and external metal surfaces. Cans produced until the mid-twentieth century, with the technique of the hot-dipping, were more resistant to corrosion than modern cans. This was not only due to a thicker container but also more specifically to a thicker (up to ten times respect to today) and more homogeneous layer of tin coating. In modern cans, the tin coating is applied with the technique of the electrolytic plating. It is often observed that these cans corrode faster, and it is therefore more interesting to study the corrosion mechanisms of their constituting material. In addition, electroplated cans

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1307

Fig. 44.6 Can presenting large stains on the label due to the perforation of the cylinder and mold growth on the label due to leakage of the content and consequent storage in an environment with high RH% (Museum Burghalde, Lenzburg)

represent the majority of cans present in collections, and this type of containers will continue enriching the museums’ collections in the future. Therefore focusing the studies of corrosion of cans on the electroplated ones is of utmost interest. Unfortunately, cans in collections are relatively rare objects, and sampling of the metal container is not possible in a non-destructive way. Therefore, the characterization of the tinplate structure and the study of the mechanisms of the tinplate corrosion were conducted on commercially available tinplate samples, fabricated according to the current standards, as surrogates of real cans. External corrosion of cans in collections is mainly due to non-adapted environmental conditions, in particular too high relative humidity, during the conservation or storage of the cans, both before and after their entry in the collection. External corrosion mostly assumes the form of rust spots appearing on the bright surface of the tin layer. This phenomenon is due to the corrosion of the base steel through gaps in the protective tin layer and the growth of iron corrosion products through holes

1308

L. Brambilla

Fig. 44.7 Can presenting large stains, embrittlement of the paper, and holes due to the leakage of the content after perforation of the metal container. In this case the loss of the label is almost total (Musée d’Ethnographie de Neuchâtel, Neuchâtel)

and pores. These rust spots can be small and appear as speckles, like in Fig. 44.9a, or cover large areas of the can, like in Fig. 44.9b. External oxidation or corrosion of the tin layer is less frequent, due to the good corrosion resistance properties of this metal. Over the 150 analyzed cans, it has been observed only in few cases, in form of blackening of the tin surfaces (see Fig. 44.9c, d). The main cause of partial or total damage of the can container is, however, internal corrosion. Internal corrosion is due to the long-term contact between the food content and the metallic container. The tin layer plays the role of a physical barrier between the iron/steel substrate and the aggressive food, and therefore it acts as corrosion protection. Nevertheless, the really long time span typical of the conservation of an object in the museums leads to the gradual dissolution of tin, resulting in the exposure of the steel to the food. This process can be more or less long, according to the quality of the tin layer, in terms of thickness and homogeneity. In fact, thick, homogeneous tin layers, such as the ones obtained by hot dipping, will

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1309

Fig. 44.8 Can with label painted directly on the metal and presenting signs of flaking and detachment due to underlying corrosion and contact with the food (Alimentarium, Musée de l’alimentation, une fondation Nestlé, Vevey)

Fig. 44.9 External corrosion on cans, taking the form of small speckles (a) or covering the whole surface (b). Cans presenting blackening of the external tin coating (c and d) (Museum Burghalde, Lenzburg)

1310

L. Brambilla

obviously resist clearly longer than thinner and porous tin layers obtained by electroplating. This is one of the main reasons why recently produced cans, present in collections, are degrading faster than older cans dating from the first half of the twentieth century. An example of the fast degradation of recently produced cans is reported here (case of the apple puree cans, reported later in Fig. 44.14). In some cases, the tin dissolution process is particularly fast due to defects already present in the tin layer at the moment of the can production. This was observed during the CANS project, for example, on low-cost cans, where SEM analysis revealed the presence of diffuse porosity in the tin layer and even the presence of holes revealing the interface tin-iron alloy underneath, as shown in Fig. 44.10. Non-adapted environmental conditions, in particular, high temperature, may also accelerate the detinning process and therefore internal corrosion. Internal corrosion may lead to two phenomena: (a) The perforation of the container. The continuous dissolution of the protective tin layer leads to the progressive corrosion of the base steel. This process will be localized in specific areas, where material defects are already present. This process will therefore lead to the perforation of the container, first through small holes, as the ones visible in the SEM image in Fig. 44.11, that will become more and more large as the corrosion proceeds, like in Fig. 44.12. In fact, once the base steel starts to corrode, the process becomes irreversible. It is self-

Fig. 44.10 SEM images of a low-quality tinplate presenting diffuse porosity and large pores exposing the tin-iron alloy present under the tin coating

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1311

Fig. 44.11 Holes on the internal surface of a heritage can due to internal corrosion (SEM image)

Fig. 44.12 Can presenting diffuse and large perforations due to internal corrosion (Alimentarium, Musée de l’alimentation, une fondation Nestlé, Vevey)

catalyzed and cannot be stopped without external intervention. The result is the leakage of the content through the holes and the possible damage of the label if the perforation is located on the cylinder, like in Fig. 44.6. In some, fortunate, cases, the perforation will occur on the bottom of the can, and the content will be drained without damaging the labels, leaving only the solid part of the food to dry inside the can. As already mentioned, the leakage of the content through holes is not only damaging the aesthetic and therefore the values of the cans, concerning both the metallic container and eventually the label. The presence of food (outside the can) in collections may attract insects and animals or lead to the development of mold, in particular if the storage environmental conditions are not adapted (too high RH%), like in Fig. 44.6.

1312

L. Brambilla

(b) The development of gas inside the can. The driving mechanisms of tin dissolution in acidic media are the hydrogen evolution reaction [8]. Therefore, the development of gas inside cans belonging to heritage collections is due to corrosion phenomena leading to the production of H2 that accumulates inside the can increasing the internal pressure. Mechanical damages may also affect the container. The further damage may be due to external shocks, like impacts, falls, or scratches. It is, however, difficult to determine if these damages were caused when the can was already part of the collection or before (even during the production phase). Mechanical damages may ruin the external tin layer and therefore promote external corrosion. Another kind of mechanical damage is the already mentioned development of gas inside the can that may cause a deformation of the metal. The container deformation due to gas development occurs mainly along the longitudinal axis, like in Fig. 44.13a, but it may occur, occasionally, also radially, like in Fig. 44.13b. During the survey, it has been possible to observe the relatively fast development of gas inside cans. An example is represented by a batch of 10 cans of apple puree produced in 2009 and included in the collection of the Burghalde Museum in Lenzburg as witnesses of the last production of cans in Switzerland by the brand Hero. Thanks to a regular survey, it was measured that all the cans already started to swell, as it can be observed in the graph in Fig. 44.14 for one can reported as an example. This increase in volume, not perceivable by simple observation and detected thanks to the regular measurements using an external caliper, is an

Fig. 44.13 Can presenting swelling along the longitudinal axis (a) and radial swelling (b) (Ortsmuseum, Küsnacht)

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1313

Fig. 44.14 Measurements performed on a can of apple puree. The measurements revealed an increase of volume of the can, not perceivable by visual observation (Museum Burghalde, Lenzburg)

indication that internal corrosion is ongoing and that this may need special surveillance and, possibly, an intervention. If the gas is not released from the can, the internal pressure may lead to the bursting of the container (see Fig. 44.15a), and the consequent projection of its content. This event is particularly undesired in museum’s collections and needs to be prevented, so as to avoid damaging other artifacts stored or exposed in proximity of the potentially bursting can, like in Fig. 44.15b. The survey conducted on the 150 cans of 5 Swiss collections allowed the identification and the classification, for the first time, of the main degradation issues occurring on food cans still retaining their original content in museums. In addition, the need to recognize and define specific indicators of cans degradation, e.g., swelling, to be monitored during periodic controls was also highlighted.

44.4

Tinplate Materials Identification and Characterization

Since their invention in the 1810s, cans have always been produced using tinplate [9] as this material presents several advantages when considering its mechanical properties and resistance to corrosion [5]. However, the production of tinplate has drastically changed in the last two centuries benefiting from the technological advances of food industry. For more than one century, the tinning of cans was performed with the traditional technique of hot dipping, allowing the deposition of

1314

L. Brambilla

Fig. 44.15 Can presenting an opening on the top due to bursting (a) (Alimentarium, Musée de l’alimentation, une fondation Nestlé, Vevey); can presenting signs of food projections due to the bursting of a can stored nearby (b) (Museum Burghalde, Lenzburg)

a thick and homogeneous layer of tin on the base iron or steel sheet, with an equal thickness on the internal and external part of the can. With the invention and the progressive introduction of the electrolytic process for tinning between the 1930s and the 1970s, the layer of tin has become thinner and thinner (from more than 13 μm in the first decades of the nineteenth century to around 2 μm in the 1930s, to arrive at a maximum of 1.3 μm nowadays). This allows to maintain similar corrosion resistance properties during the shelf life of the can, while reducing the production cost. Based on a study conducted in the 1930s by the International Tin Research and Development Council [10] on tinplate cans produced one century before, the thickness of the base steel and the tin layer within two centuries can be compared, as reported in Table 44.2. The knowledge of the material composition of the cans in collections is not only important from a historical point of view, but it is, above all, the key to the understanding of the state of conservation of a can and the ongoing of its degradation phenomena. As already mentioned, sampling possibility on heritage artifacts is limited, when not impossible, and the use of non-invasive techniques should always be privileged. However, the measurement of the thickness of a thin metallic layer, such as tin in the context of cans, is challenging, as destructive techniques should be avoided. Portable X-Ray Fluorescence Spectroscopy (p-XRF) represents a reasonable alternative to

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1315

Table 44.2 Thickness variation of tinplate cans between 1820s and nowadays Sample Roast veal can [10] Roast beef [10] Crimean can [10] Tripe can [10] Meat can [10] Currently available can [5]

Date 1824 1851 1855 1880 1938 2013

Base steel thickness (mm) 0.47 0.34 0.31 0.31 0.30 0.12–0.30

Tin coating thickness (μm) 13.47 12.32 6.93 4.62 2.31 Up to 1.35

Table 44.3 Samples of tinplate used for the calibration of p-XRF Sample 1 2 3 4 5

Nominal coating mass density [g/m2] E 1.4/1.4 E 2.8/2.8 D 5.6/2.8 D 8.4/2.8 D 11.2/2.8

Notes E: equal coating on each side of the steel D: differential coatings on each side of the steel

destructive methods. This technique presents, among others, the advantage of being fast and portable, and therefore it can be used directly in situ in museums facilities. However, XRF is a semi-quantitative technique, and a dedicated calibration of the instrument is necessary to determine the amount of tin on the surface of steel and therefore the tin thickness. For this purpose, five standard tinplate sheets, with different tin thicknesses, as reported in Table 44.3, have been investigated during the project. These tinplate thicknesses correspond to the ones currently available on the market and reported in the standard UNI EN 10202:2004 [11]. Nowadays the tinplate is produced with a differential tin coating, i.e., thickness of the tin different on the internal and external part of the can. In addition, tinplate is still produced with different tin thicknesses according to the end user (animal or human) and the different food contents (e.g., absence of additional lacquer coating for pineapple cans, as tin positively contributes as an anti-oxidant) [5]. In order to verify the nominal tin thickness, expressed in terms of mass/m2, and therefore validate the p-XRF measurements, the samples have been analyzed by using cutting-edge laboratory techniques such as Auger Electron Spectroscopy (AES) in depth profiling and chrono-potentiometry. The latter is the technique commonly used in food industry in order to determine the thickness of tin layers, following the standards [11, 12]. In addition, the roughness of the base steel substrate and of the tinplate was analyzed using a 3D White Light Interferometer. The 3D White Light Interferometry showed that the base steel presents a rough surface, characterized by parallel stripes formed during the manufacturing of the steel, as it can be observed in Fig. 44.16a.

1316

L. Brambilla

Fig. 44.16 Base steel roughness of sample 5 (a); tinplate surface roughness of sample 1 (b), with the base steel stripes still visible after tinning and 5 (c) with the base steel stripe almost no longer visible [13]

When tin is applied on the surface of the base steel by electroplating followed by the manufacturing step called flow-melting, it starts to fill the deeper zones of the base steel (valleys). If the quantity of tin applied is low, like in the case of the samples 1 and 2, for example, the tin barely covers the upper parts of the base steel (crests). Therefore, the base steel roughness is still visible on the finished tinplate, like in Fig. 44.16b. If the quantity of applied tin is larger, like in sample 5, the surface of the finished tinplate results smoother, like in Fig. 44.16c. The problem of tinplate samples presenting a low quantity of applied tin (samples 1–3) is that the average thickness on the crests is clearly lower than the total average (nominal coating mass densities) declared by the manufacturer. Therefore, in aggressive environments the tin dissolution will not be homogeneous on the whole surface of the tinplate and will be faster on the crests. The base steel on the crest will be exposed earlier to the food, making the whole tinplate less resistant to corrosion. As mentioned previously, the thickness of the tin layer on tinplate is commonly determined in food industry by chrono-potentiometry, carried out following the ASTM A630-03 2014 standard in a classic three-electrodes electrochemical cell [12]. During the measurements the pure tin, and successively the tin-iron alloy, is progressively removed from the surface of the sample until the dissolution of the base steel begins. The result of this electrochemical test is presented as a curve, like the one of sample 5 reported in Fig. 44.17, presenting three plateaus: – The plateau at the lowest potential (from 10 to 110 s) indicates the dissolution of the free tin. – The plateau at the highest potential (from 165 s to the end of the test) indicates the dissolution of the base steel. – The plateau in between (from 135 to 155 s) corresponds the dissolution of the tin-iron alloy. Subsequently, the tangents to the chrono-potentiometric curves [12] or the second derivatives [8, 13] are used to determine the thickness of the pure tin layer and of FeSn2 alloy. The thicknesses of the tin layers measured by chrono-potentiometry are reported in Table 44.4.

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1317

Fig. 44.17 Chrono-potentiometric curve of sample 5 [13]

Table 44.4 Tin thickness (μm) measured by chronopotentiometry

Sample 1 2 3 4 5

Free tin (μm) 0.215 0.211 0.426 0.936 1.061

Alloyed tin (μm) 0.124 0.106 0.138 0.049 0.193

Total tin (μm) 0.339 0.317 0.564 0.958 1.254

The thickness of the tin layer was also determined using Auger Electron Spectroscopy. AES depth profiling was obtained by scanning the electron beam (sputtering) on two areas of 20  20 μm of the tinplate. The AES sputter rate was previously determined on a sample of known thickness of pure tin deposited on copper. The determination of the sputter rate allowed to transform AES sputter times in depth. The thickness of the tin was determined taking the depth at which the tin concentration reached 50%. A typical AES depth profile for tinplate (sample 5) is reported in Fig. 44.18. The thicknesses of the tin layer obtained by AES depth profiling are reported in Table 44.5. Comparing the results of Tables 44.4 and 44.5, it is possible to observe that AES overestimates the total tin by about 15%, the thickness of the tin layer compared to the chrono-potentiometry [13]. These two techniques, even if extremely precise, cannot be applied to cans belonging to heritage collections due to their destructive nature. The results obtained by chrono-potentiometry and AES depth profiling were used afterward to verify the accuracy of the use of portable XRF as non-invasive and non-destructive technique for the determination of the tin thickness on heritage cans.

1318

L. Brambilla

Fig. 44.18 AES depth profiling of sample 5 [13]

Table 44.5 Thickness of the tin layer calculated by Sample AES depth profiling 1 analyses 2 3 4 5

Thickness in area 1 (μm) 0.544 0.382 0.622 1.136 0.907

Thickness in area 2 (μm) 0.404 0.317 0.593 1.402 1.958

Average (μm) 0.474 0.349 0.608 1.269 1.432

The correlation between the amount of tin detected by p-XRF and the thickness of the tin layer determined by AES or chrono-potentiometry is reported in Fig. 44.19. One may notice from the good correlation between the data obtained by the different techniques, that p-XRF, even with all the accuracy limitations of a portable instrument, is a promising tool to estimate the thickness of the tin layer typically present on heritage cans. In order to perform in situ, non-invasive analyses only the thickness of the external layer of tin may obviously be determined using p-XRF. However, this information, combined with historical data, may help identify the technique used for the tinning. In fact, for electrolytic tinplate, the external layer is nowadays always 2.8 g/m2; therefore the amount detected by p-XRF should be comparable to the one obtained for the sample marked 2 in Table 44.3, corresponding to recent cans. If the detected amount of tin is noticeably higher than the one obtained for sample 5 of Table 44.3, there is a high probability that the tinning was performed by hot dipping. Indeed as a matter of fact, by hot dipping it is impossible to obtain tin layers as thin as the ones obtained by electroplating. This is particularly useful to identify the

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1319

Fig. 44.19 Correlation curves obtained by plotting the Sn amount (%) measured by p-XRF vs the thickness of Sn measured by Auger (orange) and coulometry (blue), respectively [6]

Fig. 44.20 Actual tinplate structure as result of a multianalytical investigation [8, 13]

production technique for the cans produced between the introduction of the electroplating, starting from the 1930s and the obligations of indicating on the label the expiration date, which corresponds to recent regulations. This material characterization did not only allow to validate the p-XRF as portable and non-invasive technique for determining the thickness of the tin coating, but it permitted to better describe the structure and stratigraphy of the tinplate [13], which is more complex than what is usually reported in literature. The schematic representation of the tinplate in its actual structure is reported in Fig. 44.20.

1320

44.5

L. Brambilla

Corrosion of Tinplate in Complexing Acid Media Mimicking Real Cans’ Content

Tinplate cans corrode due to the interaction between the metallic container and corrosive food. The complexity of the tinplate structure and of the food chemistry makes the corrosion of tinplate a challenging interplay of phenomena to describe. As already mentioned, canned foods are extremely complex systems. They contain several molecules and ions; they cover a wide range of pH and buffering properties. Therefore, it is almost impossible to study into details the corrosion behavior of tinplate for all types of food content. However, some general conclusions can be drawn by conducting experimental studies involving compounds that are often present in canned food. Organic acids, such as tartaric, malic, citric, and oxalic are common complexing agents contained in canned fruits and vegetables. Citric acid, in particular, is commonly added to diverse types of canned foods as a pH modifier or stabilizer, due to its buffering properties. The protons provided by complexing acids, together with oxygen, are the most common oxidizing agents present in cans. Tin forms complexes with these organic acids [14, 15], and their presence reduces the activity of free tin ions in solution. According to the literature [16, 17], the internal corrosion of tinplate cans occurs according to the following routes: • The first and the last periods of the life of tinplate cans are generally characterized by rapid detinning. During the first period, the high rate of tin dissolution is due to the reduction of the residual oxygen (oxygen reduction reaction, ORR) and the oxidant species present in food [17]. In the last period, hydrogen evolution reaction (HER) is the leading process, due to the exposure of larger areas of base steel. • Normal detinning corresponds to the situation where the tin layer acts as a sacrificial anode toward the base steel. This phenomenon takes place in anaerobic conditions and in the presence of complexing agents of tin ions. The base steel may be exposed to food, through pores in the tin layer and in the FeSn2, alloy providing catalytic sites for the HER [17]. The acid corrosion of tin and the HER at the exposed steel sites occur at a slow rate. • A third corrosion mechanism may occur in anaerobic conditions and in the absence of complexing agents. These conditions correspond to a situation where tin can no longer behave as a sacrificial anode and corrosion attacks occur at the exposed iron sites. This corrosion process is usually localized and, even if not frequent, may occur in highly corrosive products (pickles and carbonated beverages containing phosphoric acid). In order to be able to predict the long-term corrosion of tinplate, aiming at determining the lifetime of cans in museums and collections, the information on the material structure and composition need to be combined with the kinetics of the corrosion reactions in the presence of food. The kinetics and mechanisms of tin corrosion reactions have therefore been studied in model electrolytes’ acidic buffer solutions of oxalic, citric, and malic acids [8].

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1321

• The concentration and pH values of these electrolytes were selected according to the values reported in the literature for can food [14, 15]. • The effect of the complexing strength change with proton concentration was investigated using solutions containing the same acid buffered at two different pH values: pH 3.0 and 4.5. • The impact of the presence of oxygen was studied using a partially de-aerated system. It is to be noted that oxygen is the oxidant present in small quantity in all cans as a result of the canning process, and independently of the kind of food. All these investigations were performed using electrochemical methods, in a classic three-electrodes cell. For the sake of simplification of the interpretation of the results, pure tin was used as working electrode instead of tinplate. This choice was supported by literature data, reporting that as long as the tinplate surface is coated with a homogeneous layer of tin, the corrosion rate of the tinplate is close to the one of pure tin [18]. A gold mesh was used as counter electrode and a saturated calomel as reference electrode (SCE). Buffer solutions of citric, L-malic, and oxalic acids were employed as electrolytes. For each buffer, two concentrations (0.1 M and 0.01 M) and two values of pH (3.0 and 4.5) were investigated. The electrochemical tests were performed in the absence of light and at low concentrations of oxygen with the aim to mimic the internal environmental conditions of food cans. The de-aeration of the solutions was achieved by bubbling argon in the electrochemical cell during the 2 h preceding the tests. Argon was then bubbled in the cell for the entire duration of the experiment in order to maintain a constant concentration of oxygen. The polarization resistance of the tin working electrode was measured by potentiodynamic polarization from
10 mV to +10 mV vs. OCP (Open Circuit Potential), using a scan rate of 2 mV/s. Cathodic and anodic branches of potentiodynamic polarization curves were recorded from OCP to
1.2 Vand
1.5 V vs. SCE, respectively. After 180 s of stabilization at OCP, the potentiodynamic polarization was recorded from OCP to 0.7 V vs. SCE, at a scan rate of 1 mV/s. A comparison between the OCP values measured in each buffer solution after the 2 h of stabilization (argon bubbling) and the corrosion potentials (Ecorr) extracted from the potentiodynamic polarization (PP) curves, reported in Table 44.6, shows that the OCP values correspond to the Ecorr. This confirms the hypothesis that, under these conditions, tin undergoes active dissolution [19] in systems mimicking tinplate cans. Moreover, one may notice that tin OCP values shift in the cathodic direction when increasing the complexing strength of the solution. Table 44.6 OCP values and corrosion potentials (Ecorr) of tin in the three buffered systems pH 3.0 3.0 4.5 4.5

[Concentration] 0.01 0.1 0.01 0.1

Citrate OCP
0.590
0.625
0.682
0.703

Ecorr
0.592
0.626
0.684
0.702

Malate OCP
0.569
0.599
0.647
0.663

Ecorr
0.573
0.601
0.650
0.662

Oxalate OCP
0.669
0.734
0.732
0.769

Ecorr
0.671
0.732
0.729
0.766

1322

L. Brambilla

The behavior of tin is similar, even if not identical, in the three organic acids; therefore, only the results obtained in citric acid, which is also the most commonly present in canned food, are reported as example in the following figures. Figure 44.21 shows the behavior of tin in citric acid in the anodic domain. The current grows exponentially with respect to the applied potential, from the corrosion potential to an overvoltage of 30–100 mV. Afterward, the currents of the active region grow linearly until reaching the passive potential (Ep). The linear evolution of the active currents with the applied potential indicates that the tin active dissolution is under ohmic control [8]. The anodic potentiodynamic polarization curves present an active region characterized by a Tafelian zone close to Ecorr, followed by the region in which the current increase with the applied potential linearly. In the cathodic domain, the system presents a short cathodic plateau in the range of potential going from Ecorr to the potential indicating the beginning of the water reduction reaction, as it can be observed in Fig. 44.22. This cathodic plateau is due to oxygen reduction based on a mechanism under diffusion control. Moving to more

Fig. 44.21 Anodic potentiodynamic polarization curves of tin in citric acid solutions [8]

Fig. 44.22 Cathodic potentiodynamic polarization curves of tin in citric acid solutions [8]

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1323

negative applied potentials (lower than
0.9/
1 V vs. SCE) after the cathodic plateau, the water reduction reaction becomes the main cathodic process. As anticipated from the literature [20], the principal cathodic reactions for buffer systems of complexing acids, mimicking tinplate cans environments, are the oxygen reduction reaction (ORR) and the hydrogen evolution reaction (HER), following Eq. 44.1 and Eq. 44.2, respectively: O2 þ 2H 2 O þ 4e
! 4OH


ð44:1Þ

2H þ þ 2e
! H 2

ð44:2Þ

Unwanted traces of oxygen in canned food are residues left during the filling process of the cans while the protons come from the acidic buffers. The same electrochemical test was performed in [0.1] citric buffer solution (pH 3), in aerated conditions, chosen as the worst, i.e., most corrosive, conditions, to evaluate the impact of residual oxygen in cans. The theoretical concentration of oxygen in these conditions, at a temperature of 25  C, is 8.3 mg/L. The results of this set of experiments, reported in Fig. 44.23, indicated that the main cathodic reaction, responsible for the presence of the plateau, is the oxygen reduction. This reaction, responsible of the fast detinning route, determines the kinetics of tin dissolution in the first period of the shelf life of a can. The kinetics of the cathodic reactions occurring in the region of potentials close to the corrosion potential (Ecorr) characterized by the presence of a plateau of current were investigated using rotating disk electrodes (RDE). The RDE allows to study the kinetics of electrochemical reactions under mixed control (charge and mass transfer). It allows to separate the contribution of the mass transport and of interface kinetics [21]. The results of these measurements, reported in Fig. 44.24, confirmed that the oxygen reduction reaction is responsible for the cathodic plateau, while the pH and the concentration do not seem to play any role justifying the presence of a plateau in the currents. Curves obtained for different concentrations of buffer solutions are

Fig. 44.23 Cathodic potentiodynamic polarization curves of tin in [0.1] citric acid solution at pH 3 in aerated conditions [8]

1324

L. Brambilla

quite similar; therefore only results obtained for [0.1] citrate solutions are reported here below. Since the potentiondynamic polarizations are independent of the rotation speed of the electrode in RDE, one may conclude that the Tafelian region of the anodic domain is not under diffusion control. As it was observed in Fig. 44.21, all anodic potentiodynamic polarization curves show an active region characterized by a limited Tafelian zone closed to the Ecorr, followed by a region in which the current increases linearly with the applied potential. The anodic Tafel slopes (ba) were extracted from the RDE potentiodynamic polarization curves at 100 rpm and reported in Table 44.7. The ba values are related to the complexing strength of the solution. Therefore, the tin dissolution in the Tafelian region is enhanced by the increase of the complexing strength of the anions present in the solution. This means that the anodic kinetics of tin dissolution become faster while increasing the complexing strength of the solution, emphasizing the importance of the knowledge of the real content of a food can to predict its long-term lifetime. This phenomenon is due to a shift to the right of the equilibrium of tin dissolution, as reported in Eq. 44.3, due to subtraction of tin ions by the complexing anions.

Fig. 44.24 Cathodic potentiodynamic polarization curves of tin in [0.1] citric acid solution recorded using RDE [8]

Table 44.7 Anodic Tafel slopes ba for tin determined on the anodic branch of the potentiodynamic polarizations of tin measured with the RDE setup in citric buffer

pH 3.0 3.0 4.5 4.5

[Concentration] 0.01 0.1 0.01 0.1

ba (mV/dec) 34.03 29.61 30.48 29.34

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

Sn⇆Sn2þ þ 2e
½19

1325

ð44:3Þ

In the RDE tests, the ORR may follow a 4-electrons process as reported in Eq. 44.1 for low rotation rates, or a 2-electrons process, as reported in Eq. 44.4, at high rotation rates, according to the Koutecky-Levich equation used to determine the kinetic parameters [21]. O2 þ H 2 O þ 2e
! H 2 O2 þ 2OH


ð44:4Þ

The theoretical limiting current values (jl) for oxygen reaction were calculated according to the Levich equation (Eq. 44.5). In the Levich equation below, the parameters are the following [21]: • F is the Faraday constant (96,485 C/mol). • c is the concentration of oxygen in the bulk (0.01 ppm). • D is the coefficient of diffusion of oxygen in pure water at 25 (2.51  10
5 cm2/s). • v is the kinematic viscosity of water at 25  C (8.93  10
3 cm2/s). • ω is the angular velocity in rad/s. j jl j ¼ 0:62nFcO2 ,b DO2 =3 ν
=6 ω =2 2

1

1



C

ð44:5Þ

In the context of the CANS project, this study revealed that a comparison between theoretical trends of jl for the ORR and the jl for the ORR measured on tin in citrate buffers suggests that for rotation rates >1500 rpm, the ORR takes place with a 2-electrons process. Further, for lower rotation rates, the ORR approaches a 4-electrons process as the high rotation rates lead to the elimination of H2O2 from the electrode surface before its further oxidation to OH
. The jl values for the ORR (jl O2/OH
) on tin in citrate buffers solutions, mimicking cans contents, have been calculated at 100 rpm and are listed in Table 44.8. jl O2/OH
can be subtracted to the cathodic branch of the corresponding potentiodynamic polarization curve in order to obtain the cathodic polarization curve of tin due only to the HER, allowing therefore to study the kinetics of the HER on tin. The cathodic polarization curve of tin in [0.1] citrate buffer solution at pH 3.0, corrected by subtracting jl O2/OH
to the cathodic branch, is reported in Fig. 44.25. The curves recorded with the RDE setup at 100 rpm approach those recorded with the static setup both for anodic potentials and high cathodic applied potentials. The intercept of the cathodic Tafel line at the corrosion potential (Ecorr) allows to determine the kinetics of tin corrosion. Even extremely low oxygen concentrations (lower than 2 ppm) increase the corrosion rates of tin up to three orders of magnitude. The kinetics of tin corrosion due only to the HER, therefore characterized by the absence of oxygen, are determined by the Tafel slope bc extracted from the corrected curve c) reported in Fig. 44.25. The bc values for the HER determined in all citrate buffers are close to 100 mV/dec independent on pH and concentration.

1326

L. Brambilla

Table 44.8 Limiting current densities for the ORR for tin in citrate buffers calculated at 100 rpm

pH 3.0 3.0 4.5 4.5

[Concentration] 0.01 0.1 0.01 0.1

Oxygen reduction reaction jl O2/OH- (μA /cm2) 0.52 0.55 0.53 0.56

Fig. 44.25 Potentiodynamic polarization of tin in [0.1] citrate buffer at pH 3.0 recorded with static setup (a) and RDE setup (b). The red curve (c) is the cathodic branch of curve (b) corrected by subtracting jl O2/OH
[8]

For the pure HER, the exchange current densities for the hydrogen (j0 H+/H2) on tin were calculated from the cathodic Tafel lines to the reversible potential of hydrogen, and the corrosion current densities (jcorr) for tin in the tested buffers were obtained by the extrapolation at the Ecorr of the cathodic Tafel line. These values are reported in Table 44.9. In the absence of oxygen, the tin jcorr corresponds to the cathodic current density of the HER on the tin surface. Therefore, the tin jcorr may be theoretically calculated using the Volmer-Butler equation [21], reported in Eq. 44.6.  jcorr,Sn ¼ jc,HER Sn ¼ j0,

þ

H =H2

exp




Ecorr,Sn
Erev,HER βc,HER

 ð44:6Þ

The tin jcorr theoretical values following the Volmer-Butler equation are reported in the last column of Table 44.9. The values have been calculated using the following parameters [21]: • Average value of bc determined in complexing acid buffer ¼ 100 mV/dec • bc ¼ 2.303 βc • j0 H+/H2 ¼ 10
4 μA/cm2, according to the literature [21]

44

Characterization and Monitoring of Complex Artefacts: Case of Food Cans

1327

Table 44.9 Exchange current density (j0) values for the hydrogen couple (H+/H2) on tin in citrate buffers calculated at 100 rpm; corrosion current density (jcorr) values obtained from the extrapolation of the cathodic Tafel lines to the Ecorr; corrosion current density (jcorr) values calculated using the Volmer-Butler equation [21] for the HER on tin pH 3.0 3.0 4.5 4.5

[Concentration] 0.01 0.1 0.01 0.1

j0 H+/H2 (μA/cm2) 7  10
5 8  10
5 9  10
5 7  10
5

jcorr (μA/cm2) measured 4  10
3 1  10
2 5  10
3 5  10
3

jcorr (μA/cm2) calculated 5  10
3 1  10
2 5  10
3 9  10
3

• Measured Ecorr,Sn • Erev,HER calculated according to the pH using Nernst’s equation The comparison between the jcorr values theoretically calculated (Table 44.9, last column) and the jcorr experimentally determined in the buffers (Table 44.9, second to last column) exhibit generally good correspondence between the values. One may therefore consider that the pH and the Ecorr of tin are the two parameters necessary to determine the kinetics of tin corrosion due only to the HER in de-aerated solution in the context of food cans. These studies conducted during the CANS project outlined the electrochemical behavior of tin in the typical corrosive environment found in canned acid products. The experiments with the rotating disk electrode, in particular, allowed to clarify the mechanisms and the kinetics of tin dissolution and to conclude that: • Tin presents an active behavior in buffered acid media containing complexing anions and low residual oxygen. • The reactions driving the rate of the active tin dissolution under cathodic control are the oxygen reduction reaction (ORR) and the hydrogen reduction reaction (HER). • The ORR is limited by the diffusion of the oxygen. • For the HER on tin, the rate of tin dissolution due to acid corrosion can be calculated using the Volmer-Butler equation. The initial concentration of oxygen present in the can after filling, the pH, and the types of complexing agents are the three fundamental food parameters for modelling the lifetime of a can. All these conclusions, in combination with the knowledge of the original content of a can as inventoried in the condition report mentioned in Sect. 44.2, and the container’s structure mentioned in Sect. 44.4, may help the conservation professionals to anticipate the after-shelf life (long-term) aging of cans in collections, further facilitating the choice of conservation strategies to be adopted. Acknowledgments This work was possible thanks to the contribution of several persons that I would like to acknowledge hereby. First of all Fabio Cova Caiazzo (EPFL) and Aline Michel

1328

L. Brambilla

(HE-Arc CR) who carried out most of the experimental work for the study of corrosion mechanisms on tinplate and the development of conservation practices for cans, respectively. Prof. Régis Bertholon (HE-Arc CR) and Dr. Stefano Mischler (EPFL) for their mentoring, advices, and collaboration over the whole duration of the project, from its onset to the conclusion, and in the context of the subsequent projects. Olivier Schinz of the MEN who conducted the survey about the presence of cans in museums. The colleagues of HES-SO Valais for their expertise in food technology. Last, but not the least, all of the Swiss museums that made their collections available for this study. All pictures of historical cans were taken by the HE-Arc Conservation-restauration team in the framework of the collaboration established with the museums cited in the present chapter (paragraph 3) for the purposes of the CANS project (2014-2017) and with the agreement of the museums themselves.

References 1. CANS project: Conservation of cAns in collectioNS – Swiss National Science Foundation grant 152946 (2014–2017) 2. Schinz O: CANS project internal report 3. Van Horn DR, Culligan H, Midgett C (eds) (2015) Basic condition reporting: a handbook, 4th edn. Rowman & Littlefield, Lanham 4. Brambilla L, Michel A, Bertholon R (2016) Condition of cans in collections: a challenge in conservation. In: Proceedings ICOM-CC metals working group conference metal 2016, ICOMCC, New Delhi, pp 266–274 5. Robertson GL (2012) Food packaging principles and practice, 3rd edn. CRC Press, Boca Raton 6. Brambilla L, Cova Caiazzo F, Michel A, Mischler S, Bertholon R (2018) Degradation of heritage cans: monitoring of museums’ collections. Measurement 127:256–263 7. Shelley M (1992) Storage of works on paper. In: Bachmann K (ed) Conservation concerns: a guide for collectors and curators. Smithsonian Books, Washington, DC, pp 29–33 8. Cova Caiazzo F (2018) Corrosion mechanisms and durability models for historical tinplate food cans. PhD thesis, EPFL 9. Peltier F, Lemoine R, Delon E (2006) La boîte, solution d’avenir. Pyramyd, Paris 10. Hoare WE, Hedges ES, Barry BTK (1965) The technology of tinplate. Edward Arnold, London 11. UNI EN 10202:2004 12. ASTM A630-03 20149 13. Cova Caiazzo F, Brambilla L, Montanari A, Mischler S (2018) Chemical and morphological characterization of commercial tinplate for food packaging. Surf Interface Anal 50:430–440 14. Sherlock JC, Britton SC (1972) Complex formation and corrosion rate for tin in fruit acids. Br Corros J 7(1):180–183 15. Willey AR (1972) Effect of tin ion complexing substances on the relative potentials of tin, steel and tin-iron alloy in pure acid and food media. Br Corros J 7(2):29–35 16. Gabe DR, Met M (1978) Principles of metal surface treatment and protection, 2nd edn. Pergamon, Oxford 17. Mannheim CH, Passy N (1982) Internal corrosion and shelf-life of food cans and methods of evaluation. Crit Rev Food Sci Nutr 17(4):371–407 18. Sherlock JC, Hancox JH, Britton SC (1972) Rate of dissolution of tin from tinplate in oxygenfree citrate solutions: I. Assessment by polarisation measurements. Br Corros J 7(9):222–226 19. Pourbaix M (1974) Atlas of electrochemical equilibria in aqueous solutions, 2nd edn. NACE, Houston 20. Bombara G, Azzeri N, Baudo G (1970) Electrochemical evaluation of the corrosion behaviour of tinplate. Corros Sci 10(1):847–856 21. Landolt D (2007) Corrosion and surface chemistry of metals. EPFL Press, Lausanne

Information and Communication Technology (ICT) for Built Cultural Heritage

45

Michela Cigola, Arturo Gallozzi, Silvia Gargaro, Leonardo Paris, and Rodolfo Maria Strollo

Contents 45.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2 The Spread of Digital Techniques in Relationship with 3D Modeling . . . . . . . . . 45.3 Data Acquisition Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.4 Elaboration of Interpretative Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.5 The Use of Information Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.6 The Context Knowledge Model for Archeological Areas Management . . . . . . . 45.7 The Case Study: The Archeological Area of Casinum . . . . . . . . . . . . . . . . . . . . . . . . . 45.8 The Representation of the Context Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.9 The Context Knowledge Structures Applied to Archeological Areas . . . . . . . . . . 45.10 First Rules’ Definition and Modeling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1330 1332 1334 1335 1337 1340 1341 1344 1345 1346 1347 1348

Abstract

The Italian territory is characterized by an extremely high number of Cultural Heritage. Their complete knowledge is extremely complex, also in relation to the multiple investigations requested. The purpose of this chapter is ICT for Built Cultural Heritage – BCH (architectural and archaeological artifacts) to collect and process the data that will be used for their analysis, safeguarding, enhancement, M. Cigola (*) · A. Gallozzi (*) · S. Gargaro DICeM – Department of Civil and Mechanical Engineering, UNICAS – University of Cassino and Southern Lazio, Cassino (Fr), Italy e-mail: [email protected]; [email protected]; [email protected] L. Paris DICEA – Department of Civil, Constructional and Environmental Engineeering, Sapienza University of Rome, Rome, Italy e-mail: [email protected] R. M. Strollo Tor Vergata University of Rome, Rome, Italy e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_45

1329

1330

M. Cigola et al.

and communication. The contribution proposes an articulated approach, aimed at the knowledge of Cultural Heritage, based on an integrated process between multiple models (computer scientist, context, information). Procedures will be defined to guarantee the quality and integration of the data acquired, providing continuous access to the information collected and processed in digital format. Most of the data will be processed as 3D digital models, incorporated into BIM systems and then processed using a Heritage BIM model. Through the creation of a data base that can also be consulted on offline and web-based systems, new forms of interaction between BCH and stakeholders will be identified and used, defining active procedures within the knowledge process. An example of this procedure will be applied to the archaeological complex of the Roman city of Casinum, in southern Lazio.

45.1

Introduction

A recent report made by the National Institute of Statistics (ISTAT) detected in Italy almost 5,000 museums and similar institutes, including public and private ones, of which 570 are monuments or monumental complexes and about 300 are archaeological areas and parks [1]. The last ones have much different characteristics each others by consistency, location, management, accessibility. Futheremore the tourist offer of archaeological interest is focused on a very small number of sites. Most of the sites scattered in our territory are in a state of neglect and are known only to experts, unknown to tourists and sometimes to the community itself. The complex issues related to the management and enhancement of archaeological parks have been the basis of a study promoted a few years ago by the Ministry of Cultural Heritage and Activities with the establishment of a work group that has developed guidelines for the definition of specific characteristics, objectives, and procedures (the guidelines for the establishment and enhancement of archaeological parks have been adopted with DM 18 April 2012, Ordinary Supplement n. 165). As defined in the document premises containing the guidelines prepared by the working group emerges clearly one of the main problems regarding numerous archaeological areas “most of which exist only on paper, not only for lack of services, instruments of management, communication systems, but primarily for the absence of a reflection on cultural premises and objectives, which should instead constitute the foundation of any initiative in this field.” With reference to the constitutional requirements – that provide for the obligation to ensure the protection and enhancement of the national cultural heritage in all its components (material and immaterial) and the need to guarantee the possibility of access and the real use by all – the reference framework related to the so-called archaeological parks is not clearly defined and in some cases subject to legal interpretations that may affect their establishment and management. It is interesting first of all to point out the difference between park and archaeological area. The first is a territorial area characterized by important archaeological evidences and by the presence of historical, landscape, or environmental values, equipped as an

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1331

open-air museum in order to facilitate reading through reasoned itineraries and teaching aids. When the archaeological component is exclusive, it is customary to use the definition of archaeological area whether it is – as in most cases – small areas with fragmentary archaeological finds, or whether there are large areas with imposing and famous remains (e.g., Pompei, Ostia, central archaeological site of Rome, etc.). The archaeological area, as defined in the Cultural Heritage Code, is “a site characterized by the presence of fossil remains or artefacts or structures of prehistoric or ancient age.” In the organization of a more detailed cataloging system, the denomination can be most articulated distinguishing, for example, between archaeological areas and archaeological interest areas, in the case in which there are potentialities of archaeological presence but already not discovered – even including the distinction for marine protected areas of naturalistic or archaeological interest or for monumental complexes. In any case it is fundamental to point out that in the processes of management and enhancement of archaeological areas and parks, it is essential to elaborate a scientific project that recognizes, analyzes, and evaluates the historical-archaeological and landscape features of the relevant area. An innovative impulse in the scientific project elaboration is closely connected to the development of digital technologies available to all operators who, for different reasons, are involved in Cultural Heritage. A broader application field – which includes the development of digital technologies – is the Information and Communication Technology (ICT) which arises linking Information Technology and Communication Science. ICT has developed and continues to develop in numerous fields, each with specific characteristics. Regarding Cultural Heritage ICT, after a first period of disorganic experimentations, is now demonstrating its numerous potential also in the processes of knowledge, protection, planning, management, and enhancement of archaeological areas and parks. However, the versatility of the computer technology underlying the ICT processes requires careful research by specialists to refine the methods for optimizing digital data management processes. The declination of ICT in the field of Cultural Heritage is characterized, perhaps more than in other sectors, by the absolute specificity of the object of investigation, especially about the semantic point of view. There is furthermore considerable economic potential considering the worth, in Italy, of the sector of Cultural Heritage. The research connected to digital technologies development – which, as said, is the basis of ICT – has followed for years two parallel paths that only recently, finally, tend to converge: on the one hand the spasmodic research linked to the continuous evolution of digital acquisition technologies (3D and 2D) and to the possibility of adapting data management software – often intentionally generic in order to be usable in more fields – and on the other hand the research on the scientific level of the product in relation to the strong interdisciplinary component that has always characterized the sector of Cultural Heritage. In other words, in many cases it has often been neglected to guarantee an adequate ratio between quantity and quality of the data, to make the process of information processing evident, to elaborate suitable communication processes without fully exploiting the potential of new digital media.

1332

45.2

M. Cigola et al.

The Spread of Digital Techniques in Relationship with 3D Modeling

The use of survey techniques and digital representation for the archaeological sites documentation is undoubtedly an important vehicle for optimizing economic resources on the preservation and enhancement of archaeological areas and parks. Moreover, it represents a great opportunity for that multitude of small sites that, as mentioned, are known to archaeologists and superintendents but very often invisible to the community and, consequently, outside the main tourist circuits. There is, quite often, a cultural short circuit of supply and demand, for which enhancement initiatives are not promoted on peripheral areas due to lack of funding which are, in turn, conditioned by the low, if not indeed non-existent, tourist demand. It would therefore be necessary to try to resolve this short circuit by implementing appropriate mediumterm strategic actions (Fig. 45.1) [2]. Digital technology is not in itself capable of solving this problem but it can undoubtedly be a strong aid. It has enormous potential but also contradictions due to both cultural differences between specialists and mutual distrust.

Fig. 45.1 The use of digital technologies in the survey of a little known archaeological site in Alba Fucens, near the ancient lake Fucino. The activity of documentation and analysis was made by Critevat (research center of Sapienza University of Rome) in accordance with the Department of Human Sciences University of Foggia, on old and new excavations. In the study of a taberna, (a) applying filters to point cloud it was possible to highlight the main finds. (b) This representation was compared with “traditional” drawing (line drawing on points cloud). (c) Sections with stratigraphy. (d) Virtual scenario of reconstruction. (By Leonardo Paris and Wissam Wahbeh with Daniela Liberatore)

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1333

Survey and representation – in all fields, not only in archeology – are disciplines that have profoundly changed in the last years of the last century. Digital has opened up new scenarios that have gradually covered numerous areas, investing all levels of intervention with different weights: from preventive investigations, to the programming and documentation of the excavation phases, from the analysis of the finds to the management by the superintendence, and from museum activities to the phase of scientific dissemination and use by the community. Digital technology, however, as mentioned, is sometimes the bearer of some contradictions: on the one hand it makes work easier, especially if one thinks, for example, of the enormous possibilities of disseminating information; on the other hand it produces strong specializations that sometimes do not dialogue with each other because it is based on different platforms and standards, such as territorial information systems, 3D modeling software, or non-invasive diagnostics, e.g., georadar or modal analysis. A specific topic about digital technologies concerns the 3D representation with the introduction of two fundamental ways of knowledge of the form (3D shape acquisition): photogrammetry and laser scanning. The first, thanks to the improvement of Structure from Motion (SfM) algorithms, makes it possible to obtain metric information adding numerous other information – typical of photo – such as wall textures, stratigraphic differences, and conservation status. The digital photography development, using some specific software, allows sometimes the realization of very realistic project prefigurations. The laser scan technology – which has higher instrumental costs but is increasingly widespread – allows to obtain, very quickly and very accurately, a 3D (by points) model. 3D modeling – more and more accessible to everyone – allows to get out of the sometimes too rigid scheme of traditional graphic models. It can return multiple views, and, through movement, it is able to propose a new perceptive dimension of the object represented in a virtual simulation of great communicative value [3]. Furthermore, the combination of photographic images and 3D models is undoubtedly the maximum result that can be obtained as a product of an integrated digital survey conducted with the most modern technologies (Fig. 45.2) [4]. As mentioned, digital technology has profoundly changed the survey methodologies as they were consolidated during the twentieth century. The development of the two 3D shape acquisition technologies has resulted in a substantial discontinuity with respect to the so-called “traditional” survey. The two phases of survey, metric acquisition and graphic restitution, previously strictly interconnected, are today very often separated. The ever-increasing use of digital acquisition techniques has progressively shifted the focus of the survey toward the data acquisition phase, moving forward, if not delegating to other professional figures, the phase of interpretation and analysis of the acquired data [5]. The differentiation of the two phases has also accentuated a problem particularly felt by archeologists concerning the legal aspects of copyright, data ownership, information sharing, and the consequent possibility of dissemination [6]. In the archaeological field, perhaps more than in other areas, this dichotomy is even more evident; more and more use is made of 3D texturized models – too often still difficult to manage both by experts in the field and by those who need to acquire information for information purposes – which very often do not have as their goal the

1334

M. Cigola et al.

Fig. 45.2 Archaeological site of Casinum. Nymphaeum Ponari, textured mesh model from points cloud. The texture of surface was obtained from a spherical panorama (on the right) in HDR, High dynamic range. (By Arturo Gallozzi, Leonardo Paris, and Wissam Wahbeh)

selection of archaeological information but the “simple” re-proposing a reality as faithful as possible to reality. It is necessary to be critical of this changed methodological framework. One wonders, for example, if these models can actually be considered as molds of reality or if in their elaboration there is some selective – and therefore interpretative – component of reality [7]. Think, for example, of points cloud quality in which the continuous, real space is discretized in a number of points with different resolutions, with the presence of voids, with photographic renderings which in turn can contain factors of inhomogeneity in setting fire or exposure. Think of the “automatic” transformation processes in numerical continuous (mesh) or mathematical (nurbs) surfaces in which “noise” reduction filters are often used, based on algorithms developed for other applications. It is therefore considered appropriate to analyze in more detail the problems related to the two phases of the survey.

45.3

Data Acquisition Phase

The first 3D model arising from the initial acquisition phase is the points cloud composed of a matrix that first of all collects the positional information of points in space in the xyz coordinates referring to a Cartesian reference system, relative or absolute. At each point usually other informations are associated such as reflectance value and color (RGB). The reflectance value depends on the amount of light energy that the detected surface “returns” to the instrument and may depend on various factors (characteristics of the material, angle of incidence of the laser beam, etc.); the RGB value derives from the integrated use of the laser technology with a camera that can be internal or external to the instrument. The laser scanner points cloud is substantially different from the photogrammetric one. This has obvious

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1335

consequences when, especially in archaeological surveys, it is necessary to use both procedures and integrate the information into unique model. Using the laser scanner it is important to carefully evaluate, depending on the object detected, the number of the capture stations, their position, the instrumental resolution, the real resolution, and the internal and external interferences. All these aspects, which constitute the indispensable basis of the acquisition project, have considerable repercussions on the relationship between quality and quantity of the final model [8]. Unlike architectural buildings in which it is always possible to recognize a specific conformation (even in the case of ruins) in the archaeological field, one often finds oneself in situations of strong uncertainty of the form and the stratifications that are its foundation. The points cloud, although very dense, fails to represent this complexity in many cases. The aspect resulting from the many experiences made of digital archaeological survey is that some fundamental information acquired by archaeologists directly on the field often does not find an equally evident confirmation in the points cloud model. This problem can be partly solved (but not always) using photogrammetric applications where the xyz position, unlike laser technology, derives from color information. The algorithms of Structure from Motion (SfM) work on the recognition of homologous points on two or more photographic images of the same context by analyzing first the chromatic characteristics of the image pixels and then defining all the photogrammetric parameters: orientation (internal, relative, and absolute), collinearity equations, and epipolar geometry. The photogrammetric points cloud quality is therefore strongly influenced by the material characteristics of the surveyed surface, much more than laser technology (unless adopting specific procedures during the acquisition phase). To all the considerations made for the laser scan technology, it is necessary to add in this case all those evaluations concerning the quality of the photographic image conditioned by numerous factors such as the distance, the characteristics of the sensor and the lens, the shooting environmental conditions, and the sequence of the different images used in the application of the photogrammetric algorithm. It is thus evident how the model resulting from the acquisition phase, consequent to the application of scientific methodologies, must be understood as a product in which the degree of correspondence with the detected object depends on the relationship between quality and quantity. So it’s always necessary to highlight choices due to environmental conditions during the acquisition phase. The lack of awareness of who, for different reasons, take part in the enhancing process of archaeological sites can lead to misunderstandings or mistakes in the following phase.

45.4

Elaboration of Interpretative Models

When the data, properly processed, returns information, it is necessary to highlight how in the archaeological field the elaboration of interpretative models takes on a specific value. The “traditional” graphic models are increasingly associated with the 3D model. Indeed in many cases the 2D graphic model derives from the 3D model which in turn derives directly from the points cloud (Fig. 45.3).

1336

M. Cigola et al.

Fig. 45.3 The Gallieno’s Mausoleum in Rome. Render view of the laser scanner points cloud (on the left) and plan with hypothesis of geometrical reconstruction (on the right). (By Leonardo Paris)

Two types of 3D models can be distinguished today: mathematical ones, in which geometric shapes are described by means of equations, and numerical or polygonal ones in which all surfaces are decomposed into simple flat forms, in most cases triangles, through identification of vertices. The difference in the elaboration of 3D models in the archaeological field (but not only) is not trivial since in the first case there is a continuous representation, in the second one a discrete type (Fig. 45.4). A points cloud, if we imagine that each point is the vertex of a triangle, can therefore be understood as a real 3D model of discrete type, as the pure and simple conversion into a mesh (i.e., into a numerical surface) of the points cloud does not change the basic information in any way. From the above, the point cloud can be defined as a numerical model by points, unlike a mesh defined as a numerical model by surfaces. There are therefore a whole series of elaborations and consequent interpretations that derive directly from the points cloud for which it is evident needing a points cloud of quality, not only with huge amount of data [9]. All the metric information deduced from a points cloud can be translated keeping the vector format (and therefore usable with other computer modeling software) or can be converted into raster mode. In the latter case it becomes even more important to carefully evaluate the final result of the survey based on the different graphical representation scales with the consequent margin of graphic error. Converting the numerical model by points in a mesh, it is necessary to carefully evaluate what has been mentioned above. From the metric point of view there is no implementation. On the contrary, it is very likely that the application of mesh surface generation algorithms results in data loss to increase software manageability at the expense of the model correspondence quality with reality. This information loss can be compensated combining color information – derived from appropriately processed photographic images – to the surface itself, with metric congruence (and the consequent accuracy value).

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1337

Fig. 45.4 Archaeological site of Populonia. Digital survey of the cistern. (a) The specific view of the points cloud, applying some filters, allows to understand the relationship between internal and external space. (b) Section. (c) Geometrical analysis of the stone ashlars. (d) 3D model (mesh). (e) 3D mathematical model reconstruction of the ashlars. (By Leonardo Paris)

This procedure is in many cases integrated in the photogrammetry processes in which, as already mentioned, the metric data derives from color, even if it is necessary to point out how the quality of the model is only partly attributable to the photographic acquisition procedures. In fact, much depends on the object detected and the level of characterization of the surfaces.

45.5

The Use of Information Platforms

Other types of instruments are now often used in addition to these models. Although research is rapidly evolving in the field of H-BIM (Heritage BIM) [10, 11], the application of the current BIM (Building Information Modeling) to archaeological heritage still requires further study. In fact, the current models do not allow to relate the context and the archaeological heritage in an effective and exhaustive way and, even if the interoperability research between the software tools GIS (Geographic Information System) and BIM are solving these problems, the research field is still poorly defined [12]. The limits are even more pronounced in the management of information relating to archaeological areas, in which it is not possible to formalize the elements before time, especially for the aspects linked to the various excavation campaigns. On the other hand, the various existing Geographic Information Systems, particularly useful for the management of spatial data, are reflected in multiple

1338

M. Cigola et al.

specialist applications – in relation to the diversified information associated with them – and can therefore be appropriately used also in the field of protection and enhancement of the archaeological site. For example, there are numerous GIS applications that, through the analysis of satellite images and historical and recent cartographies, provide a complete multicriteria analysis on archaeological areas and the relative relations with the surrounding environment(Fig. 45.5). Nevertheless, from a recent survey on the state of the art, GIS technologies still do not guarantee sufficient support in defining a complete and comprehensible framework of archaeological areas, aimed at supporting an informed planning [13]. In fact, the GIS tools allow to overlay maps, connect traditional relational databases, identify the characteristics of the urban landscape, and attach the data of the attributes in order to analyze, describe, and evaluate the archaeological areas, but do not allow to adequately reason on such information and, above all, according to artificial intelligence models. A possible integration among information systems and databases can be followed by adopting some specific procedures taken by the Context Knowledge Model (CxtKM), whose semantic structures offer multiple opportunities of analysis to verify – also in the archaeological field – data, features, problems, and intervention processes. The methodological framework is structured using a model based on ontologies that allows the interoperability with different GIS and BIM tools. The possibility of operating in different scenarios, aimed at study, protection, recovery, and intervention, analyzed by the research group, has led to the definition of a procedural practice, characterized by multiple operational aspects, which can be taken as a paradigm for similar areas intervention. A basic support to the process must involve the development of digital models through a correct representation of monumental artifacts and their environmental context. The different archaeological and architectural surveys carried out over the

Fig. 45.5 Themes and multicriteria analysis related to the archaeological area of Cassino. GIS technology. (By Michela Cigola, Arturo Gallozzi, Silvia Gargaro, and Rodolfo Maria Strollo)

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1339

years – for various reasons and with different purposes – by the research group have examined not only individual monuments, but also for the multiple relationships between them and their environmental context. This can allow the recognition and comparison of the old/new layout features and requirements, old/new spaces so as to be able to appropriately evaluate and compare various design hypotheses and validate the best solutions suited to the area of intervention (Fig. 45.6). The need for designers and operators to have an appropriate, complete, and effective knowledge of the intervention site available requires structured and complex databases, systems, and investigation procedures. Therefore, a possible operative path lies in the recourse to Context Knowledge Model (CxtKM) and to its methodological characterization in archaeological and non-archaeological contexts. In this way, it would be possible to overcome the gap that, in part, still exists between traditional research and the GIS software used to integrate surveys and information into databases that are mainly internal to the system and often with limited relational capabilities. This methodological and system approach therefore deals with how to properly combine the CxtKM model and the GIS procedures in order to define a standard protocol also for archaeological interventions on an AAKF (Archeological Areas Knowledge Framework) scheme. The latter is aimed at managing information based on the “Knowledge of Context” of archaeological sites and related monumental artifacts.

Fig. 45.6 ICT scheme. Workflow: models and applications

1340

45.6

M. Cigola et al.

The Context Knowledge Model for Archeological Areas Management

Currently the working group is developing a methodological proposal and preparing some guidelines aimed at integrating the data acquired in a GIS environment integrated by BIM procedures in the archaeological field using ontologies [14]. In particular, the research is divided into two main paths. A first aspect is addressed to the application of the criteria defined for the “Context Knowledge Model” CxtKM [15], tested on a case study in an archaeological area. The other line of application concerns, instead, specific in-depth analysis, always based on the implementation of the database of the entire study area, through GIS and BIM procedures. Using BIM applications, in the research the singular monumental artifacts were also taken into account, defining the constituting elements of the whole artifact, characterizing the three-dimensional primitives and defining corresponding ontologies. For example, through rules developed within the CxtKM, this procedure will allow to analyze appropriate monument maintenance and management strategies. The Context Knowledge Model already tested, in previous studies, can be adapted – in its fundamental procedures – to different contexts. The basic criterion of the application is aimed at facilitating the decision-making processes, taking into consideration the diversified aspects that characterize, influence, and represent the site and that – organized in tables that can be imported into the model – allow defining Bayesiantype “Inference Rules.” Figure 45.7 summarized the articulation of the components involved, in particular how the representation of the multiple information GIS can be inserted in the ontologies for the Context Knowledge. The latter can be used to model the rules for a complete and integrated understanding of the areas in question for planning, management, study, protection, redevelopment, and re-use purposes. Essentially the key components that distinguish the proposed application are identified as: – Survey phase: photogrammetric and laser scanner surveys, post-processing for data optimization, and elaboration of the models. – GIS tools: to acquire and manage information on the areas, define the constraints of the context, analyze the possibilities of re-use, and obtain new data on land management. These information organized in tables and diagrams and imported in the proposed model, through the overlapping of the databases, with the relations and rules, allow to articulate useful considerations for design purposes. – Archeological Area Invariance (AAI): the definition of the “Invariants” for an archaeological area makes possible to determine certain rules (Queries), suitable for the investigation planning (maintenance, excavation, etc.), for the definition of the possible relations among the parts of the considered area, to plan recovery operations, define visit routes, evaluate intervention strategies, etc., also useful for defining the parameters in the BIM model [16]. – Context Knowledge Model (CtxKM) allows the insertion of GIS information and ones coming from the analysis of the “Invariants,” useful to formalize the ontologies and the planning rules. Furthermore, the definition of internal and

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1341

Fig. 45.7 Outline of the Context Knowledge Model. Interrelation of GIS and “Invariants” database, definition of ontological rules. (By Michela Cigola, Arturo Gallozzi, and Silvia Gargaro)

external factors allows to evaluate the influence of the characteristics of the examined area for the evaluation of the possible intervention strategies on the area.

45.7

The Case Study: The Archeological Area of Casinum

The present work derives from a previous project that involved the archaeological area of the city of Cassino, in the lower Lazio, and developed in a coordinated research field between laboratories of several universities in Lazio: the DART, Documentation, Analysis, Architecture and Territory Survey Lab of the University of Cassino and Southern Lazio; the LAREA, Relief and Architecture Lab of the University of Rome Tor Vergata; and the CRITEVAT, Rieti research center for the environmental and territorial protection and enhancement, of the Sapienza University of Rome. In particular, during the project, a detailed survey campaign (laser scanning and photogrammetry) was conducted with the consequent 3D modeling of the main archaeological remains of ancient Roman Casinum [17, 18, 19]. Among these are the Amphitheater, the Theater, the Ummidia Quadratilla’s tomb, the Nymphaeum Ponari, as well as some significant sections of Roman roads, including a portion of the ancient Via Latina (Fig. 45.8).

1342

M. Cigola et al.

Fig. 45.8 Main archaeological finds of ancient Casinum, Theater, Amphitheater, “Tomb” of Ummidia Quadratilla, Ninfeo Ponari, Roman road. Point clouds, 3D models, graphic elaborations. Survey activities and documentation carried out by the DART Laboratories (University of Cassino and Southern Lazio), LAREA (University of Rome Tor Vergata) in agreement with the Critevat (Research Center of the Sapienza University of Rome). (By Michela Cigola, Roberto Di Maccio, Arturo Gallozzi, Leonardo Paris, Rodolfo Maria Strollo)

The morphology of the territory examined as test site for experimentation shows some interesting features and specifications: – Relations among the modern city, the archaeological area of the Roman Casinum, and the Benedictine Abbey of Montecassino – Limited extension of the area, with diversified typological characteristics of the various archaeological monumental or not artifacts present in the area – Articulated infrastructural system related to some “fragments” of Roman roads, of various consistencies – Pre-existence of paths that do not adequately integrate the various parts of the archaeological complex, with the possibility of recovering, optimizing, and systemizing the relationships of the entire plant For the aforementioned characteristics, this archaeological area is a paradigmatic example for the study of similar realities. The archaeological area is located on the edge of the contemporary town, on the slopes of Montecassino (Fig. 45.9). Among its main monumental features is the Roman theater, of the Augustan period, which lies on the hill that dominates the ancient city layout below. The urban plant is marked by the amphitheater, the largest monument in size with an imposing structure with an elliptical plan, partially set against the hillside. Among the other significant elements of ancient Casinum is certainly the so-called tomb of Ummidia Quadratilla (a Roman noblewoman belonging to the Ummidia gens), as well as some stretches of paved roads, one of which is identified with the ancient Via Latina. Slightly higher up toward the north-east, compared to the theater, in a position currently not very integrated with the rest of the archaeological area, an important typological example of the coenatio aestiva emerges in the so-called Nymphaeum Ponari: it is a portion of a large urban domus of also dating back to the Augustan age in which a rectangular room with

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1343

Fig. 45.9 The archaeological area of Cassino. Field of study for the experimentation of the Knowledge Model of Context applied to ICT

painted niches and plasterwork, covered by a barrel vault, is preserved. Furthermore, along the slope of the mountain it is possible to recognize the layout of the polygonal walls, still partially preserved, which connected the urbis to its acropolis, on the site where the Abbey currently stands. While further south, not far from the ancient urban center, there is the area of the Varronian Baths, rich in springs. Administratively, the entire archaeological area has some singularities that complicate the management and integration of the whole. In fact, the complex described is largely state-owned; on this area is located the National Archaeological Museum of Cassino, which manages part of the site, in synergy with the Archeology Superintendence of Lazio and southern Etruria to which it refers. The area of the Roman theater is the responsibility of the Municipality of Cassino, while a part of the area, on which insists the so-called Nymphaeum Ponari, is instead owned by the University of Cassino and Southern Lazio. The diversification of competences in the area represents a further aspect of interest in the elaboration of the information platform, which must be easily accessible also to the different operators and to the various professional figures involved. In this context, a first experimentation on the proposed model was done by analyzing the project of a possible highly suggestive visit route, in which the ancient Christian monument and the archaeological remains could find full integration to favor the protection, enhancement, and reuse of the archaeological complex. Also, as regards the aspects most directly linked to the enhancement, the intent is to test the procedure on a broader project that makes the different parts of the archaeological park usable, restoring importance to the visible remains and contextualizing them in the historical landscape. In this sense, a naturalistic route was hypothesized enriched with eventual structures for exhibitions and for popular and scientific activities. The project also provides teaching, educational, and informative activities, including experiences of augmented reality, to be carried out within the area (guided tours,

1344

M. Cigola et al.

educational workshops, experimental archeology, etc.), as well as defining relations with the external context, so as to interest and involve all the possible competences among the appointed institution bodies in the different administrative levels.

45.8

The Representation of the Context Knowledge

The Knowledge Representation Models allow to define information in a structured manner for each specific domain and have always been present in the CAAD (Computer-Aided Architectural Design) tools [20], but have rarely been specifically applied to Context Knowledge for interventions in archaeological areas [21]. The main languages used in the semantic web for the definition of ontologies (DAML, developed by the US Agency Defense Advanced Research Projects Agency, DARPA Agent Markup Language; OIL, Ontology Interchange Language; the BKM, Building Knowledge Management) use all the concepts introduced in the Resource Description Framework (RDF), which constitutes a structure of specifications for the description and modeling of information related to a resource, described through the definition of its characteristics or properties. In particular, the BKM theoretical model is the only one that guarantees an extended representation of knowledge in the management of a project intervention (including protection, valorization, reuse, etc.) in an original theoretical model of great usability. The entities examined in the model are defined through an MPR triplet, consisting of a semantics (Meaning), a set of properties (Properties), and a set of rules (Rules). In turn, the entities are connected to each other by semantic networks called Knowledge Structures (KS). A model is then created, consisting of entities representing knowledge, which can refer to completely different scientific fields and areas. This model, due to its constitutive connotation, has the capacity to allow a collaborative management of the process among the various professional figures involved. The conceptual basis of the model is based on the axiom that “the object of design activity is the result of a complex, dynamic and interactive system of numerous entities, each characterized by a meaning, a set of properties and a set of rules that define the relationships with the others” [22]. The proposed framework of context entities based on ontologies implements the specific Context Knowledge Model CxtKM, which has been modeled taking into account also the European INSPIRE directive and the Italian GIS agreement for GIS structured databases, all in order to allow a better interoperability with the standards and codes that are currently being defined in the field of intervention strategies in archaeological areas. The fundamental steps of the current study can be summarized in the following work phases: – Study the peculiarities of the archaeological area to be analyzed. – Detect and virtually model the artifacts on the site. – Use the GIS technologies to connect functionality, characteristics, and specificity of each artifact of the area, to verify the existing paths to valorize and the missing ones, to carry out environmental and urban analyses of the area.

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1345

– Define knowledge with information related to the context (urban, historical, cultural, economic, environmental, legal, etc.) to support designers in formulating re-use proposals and to verify them through IT procedures. – Optimize the protocols and procedures of the CxtKM model, in the Context Knowledge Process in the archaeological field.

45.9

The Context Knowledge Structures Applied to Archeological Areas

In general, each design process imposes constraints and variables whose values and parameters are context-dependent. The design activity, which is of protection, recovery, or enhancement, is therefore influenced by multiple factors, in particular by the “Context” from which the development of the project derives. Often, the changing characteristics of the environment lead to a reformulation of the requirements in the choice of possible design solutions. To overcome these difficulties, it is possible to propose Relational Structures to relate the entities involved in the archaeological area, from the morphological characteristics of the site, to the building entities, to environmental constraints, etc. This methodology allows the definition of a systematized Structure of Knowledge of Archaeological Areas (AAKF). In order to relate to each entity a definition, a specific Relation Structure (RS-AAKF) used, can be composed to solve specific objectives. In this regard, there are two fundamental classes that allow to define the Process of Information/Knowledge of Archaeological Areas: that of the Invariants of the Area, which are determined by the monumental emergencies that characterize the Identity of the Area itself and that of the Contextual Entities, peculiarities of the considered area, defined using the principles of the INSPIRE directive and the agreement for the GIS databases. In the experimentation mentioned above, relating to the planning of equipped tours of the Cassino archaeological area, the following operational phases can be identified: – Define the formal structure of the entities (meanings, properties, rules, behavior, skills, etc.). – Determining the formal models – generally mathematical – able to simulate, verify, and identify patterns of reasoning. – To use, therefore, the Context Knowledge ontologies and to apply the rules of the inferential engine and the contextual rules. It is thus possible to conceive a suitable pedestrian path equipped, which develops between the monumental artifacts of the area and which takes into account its own relational structure responding to the requirements of optimal solution [23]. The starting considerations concerned the characteristics and the relationships among the monuments of the archaeological area and its surroundings, paying particular attention to accessibility and relations with the contemporary urban context. The recognition was carried out using fillable tables in which information

1346

M. Cigola et al.

on each space and monument of the area was inserted (dimensions, materials, placement and planimetric characteristics, state of conservation, etc.). This allowed the metric/geometric definition and the historical values of buildings, spaces, structures, functional uses, materials, and their relationships. Thus the “invariants” identified which set out the criteria for operationally intervening in the area in question in relation to the archaeological/architectural artifacts present according to the defined objectives. The “Invariants” thus defined are represented by diagrams and tables in which the artifact is analyzed according to three groups of variables: Function, Properties, and Capability. Therefore, by applying this CxtKM model, through the definition of ontologies and the connection to GIS tools and databases, it is possible to support operators in the sector (not only designers but also archaeologists, site managers, administrators, etc.) to perform specific operations in their competence fields (verifying the algorithms, analyzing the constraints, checking the rules imposed at different levels, etc.). So, in brief summary, in the methodological framework, among the first operations there are the definition of the entities and the creation of rules to link entities, properties, relationships, functions, etc. Therefore the Semantic Structure of the entities constitutes the basic element of the process; this helps to provide knowledge and reduces the margins of error in decision-making operations, also allowing to identify alternative opportunities or suggestions for a wide range of analyses and interventions related to the territorial area under exam. The context can, in turn, be subdivided into sub-entities (Physical, Cultural, Normative, Economic, etc.). These can be formalized using the Protégé ontology language (OWL code – Ontology Web Language – compatible) and associating each entity rule to delineate the respective behaviors. Furthermore, it is necessary to specify that, for an appropriate definition of the rules, it is also necessary to implement the database with all the aspects related to intangible knowledge, consisting mainly of a careful review of the reference literature (historical, critical, literary, iconographic, etc.) The latter, appropriately integrated with scientific data (surveys, graphic elaborations and 3D modeling, binding regimes, technical constructive and regulatory principles, etc.), constitutes the information corpus, aimed at defining concepts and relationships that allow the organization of an adequate Structure of Knowledge.

45.10 First Rules’ Definition and Modeling Only for an illustrative aim, an example application of a simple rule is shown in Fig. 45.10, defined to verify the different hypotheses related to the walking route to realize in the archaeological area in exam. The rule identifies the possible events based on the results of the analysis: Case .a: it is not possible to use the proposed route because it does not verify the rule; Case .b: the rule is verified; Case .c: the model deduces that it is not convenient to carry out the hypothesized route, since, even if some parameters are verified (difference in height, distance from ruins, etc.) from the historical analyses made, the chosen route appears to be affected by

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1347

Fig. 45.10 Exemplification of a rule of inference of the Knowledge Model of the Context referred to a specific case, related to the hypothesized pedestrian path. (By Arturo Gallozzi and Silvia Gargaro)

possible new discoveries with consequent interference, increased costs, etc. The example shown illustrates a classic simple rule inferred from the model that is deduced from one of the other rules included in the model itself. Currently, in the case study, all the components for defining new rules, aimed at implementing the proposed model for applications in archaeological areas, are being defined.

45.11 Conclusion The experimentation, articulated on the use of multiple information platforms and tested in the archaeological case study, has more general values and aims. The choice of the Cassino archaeological area, due to its peculiarities described above, is well suited for the study in progress. It constitutes a sort of paradigm, with optimal dimensions and characteristics to define a methodological framework that can be adopted in the various similar minor sites. It should also be noted that the study group is being expanded in the course of research. In fact, diversified professional figures (archaeologists, urban planners, computer scientists) have been involved in order to optimize the project on several levels and according to multidisciplinary criteria. Consistent with the purposes of the ICT technology applications linked to the BBCC (management, study and research, diagnosis, restoration, protection, communication and dissemination, education and use). The joint work of the various experts in the field of Cultural Heritage will have positive effects on the process of

1348

M. Cigola et al.

adaptation, orientation, and expansion of technological knowledge. This process will thus be better able to respond to the specific requests in the context of the multiple cultural activities, which promote economic and sustainable development of the territories and communities, even those furthest from the main tourist routes. The system, in its diversified structure, allows to make available to users multiple products: data base (cataloging of assets, emergencies, monuments, etc.); cartography (production of technical and thematic maps, GIS calculations, risk maps, etc.); analysis (morphological and structural characterization, seismic protection, monitoring and analysis of microclimatic data, technical information, useful for multiple management, maintenance, restoration, diagnostics, etc.); artificial intelligence applications; and virtual reconstructions (virtual modeling and reconstruction, useful both for reconstructive hypotheses and for communication and dissemination, virtual and remote fruition of Cultural Heritage, etc.). Such articulated platforms, in ICT systems, allow data access and resource sharing. The enhancement of knowledge and use of cultural heritage promotes partnership with small- and mediumsized enterprises, research institutions, administrations, etc. In this view, the ultimate goal of the project is the development of a prototype, its technological experimentation, and the definition of a framework through methodological procedures and guidelines, which may find application in the various phases of the cultural heritage management process (historical study, diagnosis, protection, conservative monitoring, musealization, use, enhancement).

References 1. https://www.istat.it/it/files//2019/01/Report-Musei_2017_con_loghi.pdf. The document available on the Iast website was published on 29 January 2019 and refers to the year 2017 2. Paris L, Liberatore D, Wahbeh W (2012) Digital representation of archeological sites. recent excavation at Alba Fucens. In: Gambardella C (ed) less more architecture design landscape. La scuola di pitagora, Napoli, pp 295–304 3. Gallozzi A, Paris L, Wahbeh W (2020) 3D models and interactive communication for archaeology. In: Proceedings 2017 Computer Applications and Quantitative Methods in Archaeology (CAA) international conference, Georgia State University in Atlanta 4. Paris L, Calvano M, Nardinocchi C (2017) Web spherical panorama for cultural heritage 3D modeling. In: Ceccarelli M, Cicola M, Recinto G (eds) New activities for cultural heritage. Springer, Cham, pp 182–189 5. Paris L (2015) Shape and geometry in the integrated digital survey. In: Brusaporci S (ed) Handbook of research on emerging digital tools for architectural surveying, modeling, and representation. IGI Global, Hershey, pp 214–238 6. Arizza M, Boi V, Caravale A, Palombini A, Piergrossi A (2018) Introduzione. I Dati archeologici. Accessibilità, proprietà, disseminazione. Archeologia e Calcolatori 29:9–12 7. Palombini A (2018) Riproducibilità a vario titolo del patrimonio: situazione e prospettive. Archeologia e Calcolatori 29:87–92 8. Paris L (2010) Quantità e qualità nell'utilizzo dello scanner laser 3D per il rilievo dell’architettura, In: Vv. Aa:X Congreso Internacional Espresiòn gràfica aplicada a la edificaciòn, vol I. Editorial Marfil, Alcoy, pp 279–289 9. Paris L, Cecere C (2015) Il rilievo digitale nella cisterna del saggio IV. In: Di Cola V, Pitzalis F (eds) Materiali per Populonia, vol 11. Edizioni ETS, Pisa, pp 115–127

45

Information and Communication Technology (ICT) for Built Cultural Heritage

1349

10. Bosco A, D’Andrea A, Nuzzolo M, Zanfagna P (2019) A BIM approach for the analysis of an archaeological monument. Int Arch Photogramm Remote Sens Spat Inf Sci XLII-2/W9: 165–172 11. Gargaro S, Del Giudice M, Ruffino PA (2018) Towards a multi-functional HBIM model. SCIRES-IT-SCIentific RESearch Inf Technol 8(2):49–58 12. Dore C, Murphy M (2012) Integration of historic building information modeling and 3D GIS for recording and managing cultural heritage sites. In: 18th international conference on virtual systems and multimedia. Virtual Systems in the Information Society, Milan, pp 369–376 13. Gargaro S, Leggieri E, Fioravanti A (2014) Building heritage knowledge framework using context ontologies – regeneration of industrial areas – “Isola del Liri” case study, In: Thompson EM (ed) Fusion, Proceedings of the 32nd eCAADe Conference. Northumbria University, Newcastle upon Tyne, vol 1, pp 593–602 14. Gargaro S, Cigola M, Gallozzi A, Fioravanti A (2018) Cultural Heritage Knowledge Context. A model based on Collaborative Cultural approach. In: Kępczyńska-Walczak A, Białkowski S (eds) 36th eCAADe international conference on education and research in computer aided architectural design in Europe, vol 2. eCAADe & Lodz University of Technology, Lodz, pp 205–214 15. Gargaro S (2014) Le tecniche dell’intelligenza artificiale per la modellazione della Conoscenza del Contesto. In: Di Giuseppe E, Mazzoli C (eds) L’orizzonte del sapere tecnico in architettura. ArTec, Roma, pp 345–350 16. Apollonio FI, Gaiani M, Sun Z (2012) BIM-based modeling and data enrichment of classical architectural buildings. SCIRES-IT-SCIentific RESearch Inf Technol 2(2):41–62 17. Cigola M, Gallozzi A, Strollo RM (2016) Castrum, quod Casinum dicitur, in excelsi montis latera situm est. In: Capano F, Pascariello MI, Visone M (eds) Cirice 2016 Delli Aspetti de Paesi. Vecchi e nuovi Media per l’Immagine del Paesaggio. e-book CIRICE – Centro Interdipartimentale di Ricerca sull’Iconografia della Città Europea, Napoli 2016. Tomo 2, pp 1–10 18. Cigola M, D’Auria S, Gallozzi A, Paris L, Strollo RM (2016) The archaeological site of Casinum in Roman era. In: Bertocci S, Bini M (eds) The reasons of drawing. Gangemi, Roma, pp 201–208 19. Cicola M, Gallozzi A, Paris L, Chiavoni E (2018) Integrated methodologies for knowledge and valorisation of the Roman Casinum City. In: Matsumoto M, Uleberg E (eds) Oceans of data, 44th conference on computer applications and quantitative methods in archaeology. Archaeopress Publishing Ltd, Oxford, pp 121–134 20. Gargaro S, Fioravanti A (2013) Traditions based on context. How context ontologies can help archaeological sites. In: Future traditions, 1st eCAADe international workshop. University of Porto, Faculty of Architecture, Porto, pp 105–114 21. Simeone D, Cursi S, Toldo I, Carrara G (2014) B(H)IM – built heritage information modelling – extending BIM approach to historical and archaeological heritage representation. In: Thompson EM (ed) Fusion, Proceedings of the 32nd eCAADe conference, vol 1. Northumbria University, Newcastle upon Tyne, pp 613–622 22. Carrara G, Fioravanti A, Loffreda G, Trento A (2014) Conoscere collaborare progettare: teoria tecniche e applicazioni per la collaborazione in architettura. Gangemi, Roma 23. Fioravanti A (2008) An e-Learning environment to enhance quality in collaborative design. How to build intelligent assistants and ‘Filters’ between them. Archit Mod Inf Technol 4(5): 1–12

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage Management

46

Andrea Favretto and Bruno Callegher

Contents 46.1 46.2 46.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Relational Database, Geodatabase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . First Case. Čentur Hoard Cataloging and Management with a Relational Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.4 Second Case. A WebGIS Application Dedicated to the Ancient Coin Finds in the North-East of Italy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.5 Third Case. WebGIS Application Based on a Geodatabase (Ancient Coin Finds in the North-East of Italy) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1351 1355 1356 1358 1358 1361 1361

Abstract

In this chapter, we highlight the difference between a geodatabase and a series of GIS layers stored in a file system. After a short introduction on some database basics, we introduce the spatial component in a relational database structure. In order to clarify the explained concepts, some real cases are illustrated. They are all connected to the management of cultural heritage (ancient coins cataloging and management) with relational databases and GIS tools.

46.1

Introduction

GIS is an acronym that stands for Geographical Information System. According to each single word, a GIS is a set of related parts (System) which produces information (Information) linked to the territory (Geographical, see [13]). A plethora of GIS

A. Favretto (*) · B. Callegher Department of Humanities, University of Trieste, Trieste, Italy e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_46

1351

1352

A. Favretto and B. Callegher

definitions can be found in specific scientific bibliography. Among others, here we would like to quote the following two: 1. “The organized activity by which people: measure aspects of geographic phenomena and processes; represent these measurements, usually in the form of a computer database, to emphasize spatial themes, entities, and relationships; operate upon those representations to produce more measurements and to discover new relationships by integrating disparate sources; transform these representations to conform to other frameworks of entities and relationships.” [9] 2. “A system that contains spatially referenced data that can be analyzed and converted to information for a specific set of purposes, or application.” [1]

The first one focuses on considering GIS mainly a tool for analyzing spatial data, while the second one stresses that GIS can be considered basically an information system. “This means that a GIS collects data, sifts and sort them and selects and rebuilds them to find the right information to answer a question” [10]. As regards to GIS data, it is well known that they can be divided into spatial and attribute data (also called “feature codes” – [5]). Spatial data are managed with vector and raster structures, while attribute data are stored in tables connected to the spatial data. Attribute tables can also be connected to each other by relational database rules, using SQL (Structured Query Language) commands like “join” or “select” (to deepen the SQL and Relational Database methodology, see among others [12]). As stated, attribute tables (connected to the GIS layers) can be joined to other tables using SQL commands. The classification in class of intervals of the SQL joined fields is a common way to build thematic maps that illustrate many geographical analyses. In these studies, the term “Database” is sometimes misused. The attribute tables of the vector layers are often called “Database” even if they are not connected with each other, and they are only saved in the file system of the computer. A relational database is quite different. Its origin goes back to libraries, civil registries, and medical and firm records, even further the IT era [2]. In 1970 Codd [11] created the so-called relational model of a database system. This model was built in order to keep the consistency of the data stored in the database, avoiding the redundancy in the information base. Consistency is one of the four rules that ensure accuracy of all the database transactions (ACID). The acronym ACID stands for Atomicity (transactions are atomic – “all or nothing” rule), Consistency (transactions transform the database data from a consistent state to another consistent state), Isolation (transactions are isolated from one another – they are securely and independently processed without interference), Durability (transactions are completed in that once committed their generated changes are stored permanently). See Date, op. cit., [3]).) A database system is essentially an archive in which all the digital data are integrated and shared. This target is achieved by the relationships among the database tables (based on the equality in the content of the primary and foreign key fields).

46

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage. . .

1353

This mistake is partially justified by the closeness of GIS and Geodatabase concepts. Their differences are gradual. If we consider only the data component of a Geographic Information System, Geodatabase and GIS are equivalent. As we wrote, a GIS can be seen both as a tool and an information system [1]. If we think of a GIS mainly like an information system, the Geodatabase concept can be considered as its synonym. The main issue is data integration. (The data integration in a relational database means that there is no duplication of information in the database structure.) A relational database is a data integrated structure. Data integration is achieved by constructing the relationships between tables. If many attribute tables are inordinately stored in the file system (even if these tables are connected to some spatial layers), this is not a geodatabase. In this case, you only have a collection of GIS layers (with their attribute tables) stored in your hard drive, whereas a geodatabase is a relational structure. In the last years GIS methods became a widely adopted instrument by operators in the Cultural Heritage (CH) field (see [20], for a report about the use of GIS in CH). GIS technology was adopted by several Public Authorities responsible for CH in order to create complex and sometimes integrated information systems. Browsing the Petrescu report [20], it is clear that GIS technology is considered an implemented tool on a certain data model (often not precisely stated), in each different quoted country. Many other papers, on the contrary, describe more or less thoroughly the use of GIS in CH. They present many information system structures with a spatial component, connected to GIS features by means of SQL commands. In these papers, sometimes we can find a certain inaccuracy of terms because they often speak of GIS, Geodatabase, SIS (Spatial Information System), Spatial GIS, Geo-DataBase Management System, and spatial information. Sometimes these terms are used as synonymous, but in scientific literature they have a precise meaning, which is linked to their respective features. Vacca et al. [24] present a SIS for the Architectural and Cultural Heritage of Sardinia. The SIS is “composed by a geodatabase, a GIS and a WebGIS.” The geodatabase structure (built with the open source PostgreSQL DBMS – DataBase Management System) has a spatial component (built with the PostGIS extension), which stores the location of each architectural unit in the database. (“This extension turns the DBMS into a Geodatabase by adding functions and data types that can manage geometries and reference systems” [24]). GIS functions are used to access the database using SQL commands in the QGIS open source software environment, while the WebGIS is built using Leaflet Javascript open libraries. Malinverni et al. [18] use a database structure to store the data related to the Villa Buonaccorsi historical garden (Potenza Picena Municipality, Macerata Province, center of Italy). They speak about a geodatabase “structured like a Relational Data Base Management System (RDBMS), allowing the data management on the relationship model proposed by Codd.” The logical structure of the tables is presented and then is specified that “the tables will be linked to the georeferenced vector data (points, polylines and polygons) which represent the single objects in the Garden,

1354

A. Favretto and B. Callegher

identifying the correct relationship by means a univocal identification (ID) code.” The georeferenced vector data represent the natural and artificial elements inside the garden and are managed by shapefiles (shp) in the QGIS software environment. The attributes of each shp file is assigned using the SQL join command in QGIS. In this way a destination derived table is built. This table “receives the information derived from another table (source).” This solution is quite different from the previous instance (Vacca and others), because the spatial component is stored outside the database structure and then is linked to the database tables. A geodatabase should instead be a structure with a spatial component stored inside (generally by the spatial extension of the DBMS software – PostGIS in the PostgreSQL case). Guarnieri et al. [16] built a geodatabase structure to manage multisource data (raster and vector layers, text, videos, images, etc.). The study site is villa RevedinBolasco and its park in the Castelfranco Veneto Municipality (Veneto RegionNorthern Italy). The geodatabase contains spatial information (remote sensed georeferenced images at different spatial resolution together with vector floorplans and placemarks). The same uses PostgreSQL as DBMS (with its spatial extension PostGIS) and allows the download of dataset in specific sections of its WebGIS platform (built with HTML5 and PHP programming languages). Another paper [21] speaks about a “simple data structure” in order to allow an easy access to the information about the Fanum Fortunae Roman forum (Fano Municipality, Pesaro and Urbino Province, Marche Region, Italy). The data structure has four so-called “key aspects” (we quote here only the first three, because the fourth is connected to 3D visualization and virtual reality). They are: definition of the database structure, development of a spatial database management system (Geo-DBMS), spatial data handling into a GIS environment. The database structure (“physical implementation of tables, fields, relations and queries”) was made using Microsoft Access, while an open source GIS platform (QGIS) was used to achieve the further steps of the project. The authors specify that “the Geodatabase is a collection of spatial data of various kinds, such as the municipal mapping, orthophotos, DTM and vector files arising from topographic surveys, laser scanners and performed for this project, as well as the entities associated to the relational database created in Access environment.” The relational structure is the MS Access one, which is later connected to the spatial component of the database (the Geo-DBMS in the key aspects) using the SQL functionalities of QGIS software. When speaking of CH and GIS, we can remember the Concept Reference Model (CRM) developed by the International Committee for Documentation (CIDOC) of the International Council of Museums (ICOM). CIDOC CRM “is a theoretical and practical tool for information integration in the field of cultural heritage” (see http:// www.cidoc-crm.org/). As one of the CIDOC CRM objectives is the “information exchange and integration between heterogeneous sources of CH information” (see Definition of the CIDOC Conceptual Reference Model – http://www.cidoc-crm.org/ html/5.0.4/cidoc-crm.html#_Toc310250685), this model is often used as the basis for the schema of a CH database. CIDOC CRM was implemented in a relational

46

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage. . .

1355

database (see the Musinfo project – [19]) and also connected to GIS to do spatial visualization or analysis [17]. This was made possible by the capability of the relational model to store heterogeneous data inside its structure. If the different entities and relationships are well recognized and represented in the database schema, the built IT application can manage efficiently the data for a very long time (at least until new entities need to be included in the database schema and/or new relationships need to be represented). The relational model often causes very complex schemas to be built. This happens especially when the data structure is complex too. This can bring on difficulties when the data structure changes; therefore the schema changes too and all the registered data must be fitted in the new schema. The cost of this schema updates caused the implementation of the so-called NO SQL database. No SQL dbases are a possible solution to the unstructured exponential growth of data that is typical of the Web environment. They can therefore be found in companies like Facebook, Google, and Amazon, which have to face large volumes of data with a structure that is constantly changing (to deepen NO SQL dbases, see, among others, [15]). In this paper, we highlight the difference between a geodatabase and a series of GIS layers stored in a file system. After some database basics, we introduce the spatial component in a relational database structure. In order to clarify the explained concepts, some real cases are illustrated. They are all connected to the management of cultural heritage (ancient coins cataloging and management) with relational databases and GIS tools. First, a relational database of ancient coins with no spatial component is presented. The case of a WebGIS application dedicated to some ancient coin finds in the North-East of Italy follows. Another WebGIS application, based on a Geodatabase that manages the same data, is finally shown, in order to highlight the differences between the two ways of storing and managing coin data.

46.2

Relational Database, Geodatabase

A database can be defined as a digital archive. Date [12] speaks about a “computerized record-keeping system,” while Oracle corporation, one of the main database software producer, refers to “a collection of data treated as a unit” (https://docs. oracle.com/cd/B19306_01/server.102/b14220/intro.htm). A layer of software usually handles all the user requests and it is called DBMS. Database management consists in a series of operations, including: • • • • • •

Adding new files to the database Inserting new data into the added files Updating data in stored files Displaying data from stored files Deleting data from stored files Removing one or more stored files

1356

A. Favretto and B. Callegher

Data are stored (and perceived by all the users) as tables and nothing but tables. Tables are usually non-square matrices, consisting of columns (fields) and rows (records). Data are integrated and shared in a database. This leads to the major advantage of the whole methodology and means that there is no redundancy among the stored files. If a particular article of data is recorded only once in the archive, the same has to be shared, and, in this way, all the information in the database is managed more efficiently. For instance, updating procedures are faster. Above all, data integration ensures the integrity of the database. Integrity refers to the accuracy and correctness of data in the database. Data integration is achieved by relationships (also called “associations”) among the database tables. These relationships can be “one to one,” “one to many,” and “many to many,” according to the characteristics of the entities about which the information is needed. An entity is “any distinguishable object that is to be represented in the database” [12]. Relationships are based on the content equality of two fields belonging to two different tables. These common fields are called keys. Keys can be primary (PK – is a unique identifier of a given record in a table) or secondary (also called foreign key – FK – a field in one table that refers to the PK in another table). PK and FK are accordingly used to link two tables together in the form of a “one to many” relationship (see [12], or the SQL online tutorial – https://www.w3schools.com/sql/ – for further information of the other relationships mentioned above). When PK and FK values are equal, the records of the two tables are linked. Database tables can also contain spatial geometries (both vector and raster). In this case the database is called Geodatabase (or also “spatial database”; see [22, 23]). The spatial extension that manages the stored geometries enriches the database with GIS functionalities like spatial queries. In this way the DBMS becomes a real Spatial DBMS which can query and elaborate spatial data and can be connected to any GIS software.

46.3

First Case. Čentur Hoard Cataloging and Management with a Relational Database

The treasure found in Čentur (Slovenia), consisting of many thousands of coins (mainly folles), is a well-known case in the numismatic literature [4, 7, 8]. In fact, the total amount of folles was published in successive stages, between 1934 and 1962, as a result of a long series of finds that have been carried out over more than 20 years, all in the same site. Recently, some researchers have hypothesized that the discovery was one single discovery and not a series occurring over a long period of time, that it happened in one period of time, but that it was disclosed over many years mainly for reasons related to the geopolitical situation along the border between Italy and Yugoslavia. Today, the Čentur treasure is held by a small number of Institutions/Collectors, in more than one country. In order to catalog and manage the information of the whole treasure, a centralized relational database was planned. The database is managed by a

46

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage. . .

1357

Fig. 46.1 The Čentur hoard database structure

client/server architecture, available online by a limited number of clients (the owners of the Čentur treasure). The database structure is shown in Fig. 46.1. In total, there are seven tables: 1. “coin” table contains information about the coins. The table is connected to the “initial_owner,” “final_owner,” and “ric” tables by three “one to many” relationships (FK in the coin table and PK in the other tables). The same table is also connected to “mint” and “authority” tables by a “many to many” relationship (using the join table named “connect_mint_aut_coin”). 2. “initial_owner” table contains information about the first owner of the coin. 3. “final_owner” table contains information about the current owner of the coin. 4. “ric” table contains information about the RIC volumes that give the coin references. Roman Imperial Coinage (RIC) is the British catalog of the Roman coin minted from 31 BC to the end of the Roman Empire. 5. “mint” table contains information about the different mints that minted the coins. 6. “authority” table contains references about the authorities that issued the coins. 7. “connect_mint_aut_coin” table is a join table that connects “coin,” “mint,” and “authority” tables. We developed the database structure using PostgreSQL, a powerful open source relational DBMS, with over 30 years of active development (see https://www. postgresql.org). In order to simplify the administration and development of the database, we also used pgAdmin. This is an open source GUI (Graphical User Interface) that simplifies the creation, maintenance, and use of the database objects (see https://www.pgadmin.org). Looking at the database structure, it’s easy to realize that no spatial component was planned. This is because the database catalogs information that comes from the same place (Čentur). No spatial query was considered necessary in the database

1358

A. Favretto and B. Callegher

structure development phase of the project. PostgreSQL, however, has an efficient spatial extension, named PostGIS, that can be applied at any time to the database in case of a future development in this direction.

46.4

Second Case. A WebGIS Application Dedicated to the Ancient Coin Finds in the North-East of Italy

Using a catalog of ancient coin finds in the Trieste and Gorizia Provinces (North-East of Italy), taken from Callegher [6], a WebGIS application was built. The source catalog contains the territorial distribution of Celtic, Roman (republican and imperial), Byzantine, and Longobard coin finds in the provinces of Trieste and Gorizia (Friuli Venezia Giulia Region, Italy). The application can be found on the Website named “Cartografia dei ritrovamenti monetali di età romana in Friuli Venezia Giulia” (http://disugis.units.it/Numismatica/Localita/index.html). The interactive maps on the Web were realized using OpenLayers, an open source software library to “further the use of geographic information of all kinds” (https://openlayers.org). All the coin finds were georeferenced (not in their exact location, according to the current Italian laws on protection of archeological heritage) and can be interactively queried by the users. We chose OpenStreetMap as base map of our application for its Creative Commons Attribution-ShareAlike 2.0 license (see https:// creativecommons.org/licenses/by-sa/2.0/). The users can only make graphic queries to the Website, placing the cursor on the different coin symbols. The given coin information is: • The location of the coin finds on a remote sensed image (an ortho-image granted by the Friuli Venezia Giulia Region). • A photograph of each coin find place. • All the find information (in pdf format), taken from the Callegher catalog. Figure 46.2 shows the home page of the WebGIS application. Gorizia Province waypoints are highlighted in red while Trieste Province ones are in blue. Of course, all waypoints are interactive. Because of the fact that the coins are grouped by their find location, we did not build a database structure. All the GIS vector layers (with their attribute tables) are stored in the file system of the Web server, together with all the html files needed for the functioning of the application.

46.5

Third Case. WebGIS Application Based on a Geodatabase (Ancient Coin Finds in the North-East of Italy)

We present another WebGIS application, this based time on a Geodatabase. We used the same data of the previous application (the ancient coin finds in Trieste and Gorizia Provinces). Instead of starting from the find place, we considered the single

46

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage. . .

1359

Fig. 46.2 The home page of the WebGIS application dedicated to the ancient coin finds in the Trieste and Gorizia Provinces (North-East of Italy). See http://disugis.units.it/Numismatica/ Localita/index.html

coins, in order to query the database both graphically and for some of its attributes (for instance, the mint that minted them or the authority that issued the coins). The Web application can be seen at http://disugis.units.it/ritrovamenti%20monetali% 20epoca%20romana/index.html. Figure 46.3 shows the home page of the application. On the top right hand side of the figure, there is a window with some available attribute queries. They are “mint,” “nominale” (denomination of the coin), and “period”. (We would like to note that these three attribute queries are not the only possible attribute queries to the database. But these are the queries we chose to insert in the WebGIS application.) The default base map of the application is by Pelagius project (http:// commons.pelagios.org), granted under a Create-Commons 3.0 (CC-BY) license. OpenStreetMap base map is also available. We used Leaflet open-source JavaScript library to build the application (https://leafletjs.com). For a more in-depth explanation of the WebGIS application and the ancient coin catalog, please see [14]. As already stated, the WebGIS application is based on a relational database with a spatial component (Geodatabase). Figure 46.4 shows the database structure. Six tables make up the database, related by five “one to many” relationships. The relationships among the tables are built by primary (PK) and foreign keys (FK). The six tables are: 1. “monete” (coins) contains information on the coins. 2. “zecca” (mint) contains information about the mints that minted the coins.

1360

A. Favretto and B. Callegher

Fig. 46.3 The home page of the WebGIS application dedicated to the ancient coin finds in Trieste and Gorizia Provinces (North-East of Italy), based on a Geodatabase structure. See http://disugis. units.it/ritrovamenti%20monetali%20epoca%20romana/index.html

Fig. 46.4 The Trieste and Gorizia Provinces ancient coin finds Geodatabase structure

3. “tipo_moneta” (type of the coin) contains information about the denomination (type) of the ancient coin. 4. “coordinate” (coordinates) contains all the information on the locations at which the coins were found. This table contains the spatial component of the database (latitude/longitude fields – see Fig. 46.4). 5. “autorita” (authority) contains references to the authority that issued the coin. 6. “periodo” (period) table contains only one field. This table is used for querying purposes by the WebGIS application. The database structure was built using PostgreSQL with PostGIS extension, the spatial objects official provider for PostgreSQL (https://postgis.net ).

46

Relational Database, GIS Layers, and Geodatabase for Cultural Heritage. . .

46.6

1361

Conclusions

We have attempted to clarify a concept that is sometimes not fully understood by GIS insiders. It is agreed that the attribute tables of a GIS layers collection (inordinately stored in the file system) cannot be considered a relational database. However, the question is: is it completely correct to consider a set of GIS files (graphic and attribute ones) a GIS? Indeed, if a GIS is considered as mainly an information system (a spatial database) and not as a working instrument (a set of geographic functionalities), maybe the answer to this question is “no.” This is not the suitable context to philosophize over GIS (in IT no one changes what works), although, at a less inflexible level, we can say that there is a big difference between a geodatabase and a set of GIS layers stored in the file system. A geodatabase is a relational database with a spatial component. A set of GIS layers stored in the file system also has a spatial component. However, GIS layers lack the so-called data integration (an important specific feature of a database). As we have seen, this feature is achieved creating relationships among the tables. The key to understanding the core of this issue is precisely the presence (or absence) of the relationships among the tables in the database structure. Only if tables are related to each other, we can speak of geodatabase; if not, we only have a set of GIS files stored in the computer mass memory.

References 1. Antenucci JC, Brown K, Croswell PL, Kevany MJ, Archer H (1991) Geographic information systems. A guide to the technology. Chapman & Hall, New York 2. Berg KL, Seymour T, Richa G (2013) History of databases. Int J Manag Inf Syst 17(1):29–36 3. Bernstein PA, Newcomer E (2009) Principles of transaction processing. Morgan KaufmannElsevier, Burlington 4. Brusin G (1937) Notiziario archeologico (1935–1936), Atti e memorie della Società istriana di archeologia, vol 47 5. Burrough PA (1986) Principles of geographical information systems for land resources assessment. Clarendon Press, Oxford 6. Callegher B (2010) Ritrovamenti monetali di età romana nel Friuli Venezia Giulia. Province di Gorizia e Trieste, RMRFVG III-IV. EUT, Trieste 7. Callegher B (2015) Un milione di denari sulla collina di Centur. In: Garaffo S, Mazza M (eds) Il tesoro di Misurata (Libia). Istituto Nazionale di Studi romani, Roma 8. Callegher B, Favretto A (2019) Thousands of Tetrarchy folles all over the world: an hypothesis of re-composition. In: Too big to study?, Polymnia. Numismatica Antica e Medievale. Studi, vol 11. EUT Edizioni Università di Trieste, Trieste 9. Chrisman N (1997) Exploring geographical information systems. Wiley, New York 10. Clarke KC (1999) Getting started with Geographical Information Systems. Prentice Hall, Upper Saddle River 11. Codd T (1970) A relational model of data for large shared data banks. Commun ACM 13(6):377–387 12. Date CJ (1994) An introduction to database systems. Addison Wesley, New York 13. Favretto A (2006) Strumenti per l’analisi geografica. GIS e Telerilevamento. Patron, Bologna

1362

A. Favretto and B. Callegher

14. Favretto A, Callegher B (2017) Burgon’s expectation: ancient and new cartographic visualization for numismatic data and coin finds. Cartographica 52:2 15. Favretto A, Montagnari M, Vidulli M, Moser S (2019) Database relazionali e “NoSQL”. Alcuni spunt per un corretto inquadramento delle metodologie. ASITA 16. Guarnieri A, Masiero A, Piragnolo M, Pirotti F, Vettore A (2016) A geodatabase for multisource data applied to cultural heritage: the case study of Villa Revedin Bolasco, International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLI-B5 17. Hiebel G (2010) A relational database structure and user interface for the CIDOC CRM with GIS integration. 22nd CIDOC CRM and 16th FRBR CRM 18. Malinverni ES, Chiappini S, Pierdicca R (2019) A geodatabase for multisource data management applied to cultural heritage: the case study of Villa Buonaccorsi’s Historical Garden, International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XLII-2/W11 19. MCN (1999) Implementing the CIDOC CRM with a relational database. Spectra 24(1):1–6 20. Petrescu F (2007) The use of GIS technology in cultural heritage. In: XXI international CIPA symposium 21. Pierdicca R, Malinverni ES, Tassetti AN, Mancini A, Bozzi CA, Clini P, Nespeca R (2015) Development of a GIS environment for archaeological multipurpose applications: the Fano historic centre, Heritage and Technology, XIII International Forum, Le Vie dei Mercanti 22. Rigaux P, Scholl M, Voisard A (2001) Spatial databases: with application to GIS. Morgan Kaufmann Publishers, San Francisco 23. Shekhar S, Chawla S (2003) Spatial databases: a tour. Pearson Education Inc, Prentice Hall 24. Vacca G, Fiorino DR, Pili D (2018) A spatial information system (SIS) for the architectural and cultural heritage of Sardinia (Italy). ISPRS Int J Geo-Inf 7:49

3D Digitization of Tangible Heritage

47

George Pavlidis and Anestis Koutsoudis

Contents 47.1 47.2 47.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3D Digitization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3D Digitization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.1 Short-Range Laser Beam Triangulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.2 Shape from Silhouette . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.3 Shape from Stereo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.4 Shape from Structured Light . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.5 Shape from Shading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.6 Shape from Photometry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.7 Shape from Texture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.8 Depth from Focus/Defocus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.9 Shape from Shadow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3.10 Structure from Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4 The Influence of Object Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4.1 Object Size and 3D Digitization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1364 1365 1367 1368 1371 1374 1377 1380 1381 1384 1384 1389 1391 1394 1395 1399 1400

Abstract

Digitization of tangible entities is the process that focuses on the transformation of the real world and its features to a virtual world, the digital world of computing devices. This virtual world comes along with a typical set of rules, benefits, limitations, and opportunities. As this world is mathematically digital, every virtual entity is discretized and quantized, being only an approximation of its real counterpart, or “source,” or “parent.” Digitized tangible entities run a parallel life, potentially a life with no ending, as the most apparent benefit of G. Pavlidis (*) · A. Koutsoudis (*) Athena-Research and Innovation Centre in Information, Communication and Knowledge Technologies, University Campus at Kimmeria, Xanthi, Greece e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_47

1363

1364

G. Pavlidis and A. Koutsoudis

the virtual world is the long-lasting presence. Since tangible heritage entities constitute the most radiant examples of human civilization, they are the obvious targets for digitization. Particularly, since these entities are of three dimensions (3D), their digitization should also follow the 3D world representation; let us forget the fourth dimension, time, in this case, since this is the dimension that can be “frozen” with 3D digitization. 3D digitization of precious entities not only safeguards them in the virtual world but also opens up new horizons for presentation, knowledge dissemination, research and study, conservation, and even physical duplication. In order for these horizons to actually open, human ingenuity with long-lasting and painstaking research and development resulted a number of 3D digitization methods, all targeting the best possible result in terms of an accurate virtual replication. This is the focus of this chapter, in which an attempt is made to explain the ideas and inner workings of the various 3D digitization methods that have been invented during the last half of the twentieth century and the beginning of the twenty-first.

47.1

Introduction

Formally, digitization is the recording of quantized values of discrete measurements of physical quantities. In essence, it is the conversion of analog quantities of the physical world into digital values in a computer system. A digital quantity is nothing more than a set of rounded values of a systematically sampled physical quantity. Although in typical computing systems the representation of digital quantities is in binary form, it is emphasized that digitization is not connected with the creation of binary representations. Thus, for example, the conversion of a set of temperature measurements taken at 1 min intervals into integer values (say, in integer
C) is a digitization process. Particularly for tangible cultural heritage, digitization can be defined as the process of digital recording of artifacts, structures, monuments, manuscripts, and any object that conveys information about history, tradition, language, art, religion, and culture, including science and technology. The motivation for the creation of digital copies is multifaceted. Study, research, comparative assessment, preservation, safeguarding, dissemination, education, and duplication are among the most pronounced motivations. When digitization focuses on the recording of the geometric structure and the spectral signature of three-dimensional (3D) physical objects, it is termed 3D digitization. The main goal of this chapter is to provide a concrete introduction to 3D digitization, particularly in the domain of tangible heritage, in which special requirements, specifications, and restrictions apply. The text concentrates and highlights the basic ideas on the principles of the various digitization methods, the way they are implemented, and the needs and challenges that arise in the highly demanding applications in cultural heritage.

47

3D Digitization of Tangible Heritage

47.2

1365

3D Digitization

Although two-dimensional (2D) digitization (typically known as “scanning”) is considered a rather simple process, creating digital copies of real-world 3D objects is a challenge, especially when their geometric complexity goes beyond elementary solid shapes, such as spheres, pyramids, and cubes. As 3D digitization, in its typical and most widespread use, achieves the geometric and spectral recording of objects’ surfaces, it results in the creation of 3D empty shells. Thus, apart from tomographic methods, the vast majority of 3D digitization methods are unable to record information about the internal structure of the objects and in some cases it is said that the 3D replicas are actually of 2.5 dimensions (2.5D). In addition, as most methods depend on the modulation, manipulation, and detection of electromagnetic waves, like visible or infrared light, 3D digitization methods are usually references as optical documentation methods. Last but not least, following the terminology for 2D digitization, some particular 3D digitization methods that follow the scanning paradigm in 2D scanners are called 3D scanning methods. In practice, 3D digitization consists of the measurement of the geometric structure and the spectral response of an object’s surfaces, usually point by point. In order for 3D digitization to be complete, those measurements should be carried on the entire object surface, and to be as dense and accurate as possible. 3D digitization can be based on a wide variety of methods and systems, and, thus, there are significant variations in the applications, the practice, the workflow, and the outcomes. In general, the systems can be categorized to mobile, portable, and immovable or fixed. Of those systems, there are those that depend on highly specialized and expensive equipment, whereas others depend mainly on algorithms and software solutions. Some of the available methods require a long on-site period, or others require more time at the laboratory. Some methods are based on visible light detection, others on infrared, others on controlled lighting, and others on touch with the objects. In the past, 3D digitization approaches were not efficient, mainly due to the lack of specialized measurement devices, powerful computing systems, and sufficient storage space. Today, these tools are available and have led to an evolving market for 3D digitization systems. These systems allow the efficient recording of morphological (geometric) and spectral (color) information. Their principle of operation is based, among other things, on the implementation of various mathematical models and smart ideas. Nevertheless, it is still certain that each system is based on a series of assumptions and is subject to specific constraints. The interest of the research community focuses on the development of algorithms for the efficient data processing and on the creation of new accurate optical recording approaches and methods. It can be said with certainty that the development of 3D digitization systems is the result of collaborative work among scientists from different fields of research, such as computer vision, 3D graphics, engineering, measurement theory and technology, of photogrammetry, and, of course, mathematics.

1366

G. Pavlidis and A. Koutsoudis

An important role in increasing the popularity of 3D digitization methods played the dramatic improvement in 3D real-time graphics systems and methods. The ability to manage complex 3D geometries at low cost is directly applicable to the display of detailed, highly accurate 3D models, which are usually derived from 3D digitization systems. Nevertheless, to date, only a very small number of methods approach the idea of scanning at the touch of a button, and increasing the level of automation is still a topic of research and development. The domain of tangible heritage is rather challenging for 3D digitization methods, due to the wide range of object types and sizes, the diversity of materials, and the significant variation in geometric complexity. Nevertheless, 3D digitization is today a particularly important tool for optical documentation of cultural heritage. Certainly, the complete digital documentation of cultural objects should provide an ontologically complete picture (the term “ontology” in this regard relates to its use in the field of information technology, where it is linked to the semantics of a “digital entity”) of the objects for the future scholars and laymen. After decades of research and development, adaptations, and simplifications, 3D digitization is now a common practice in the field of cultural heritage research. Some of the most important benefits it offers include: • Preservation of a sensitive, aging, and threatened wealth. 3D digitization facilitates the study of possible reconstructions and the experimentation on digital copies instead of the artifacts themselves. Since it creates digital snapshots of the status of cultural objects, it works like a time capsule and is a valuable tool for preservation. • A range of options for the experts. 3D digital copies can be used to populate digital repositories with capabilities to automatically categorize, retrieve, and view specific subsets of the repository, possibly revealing relations that were not previously identifiable. 3D digital copies can also be used to create physical copies through 3D printing technologies for various purposes, including reconstruction, maintenance, accessibility, and exhibition. In addition, the use of 3D digital copies on the Web can provide simultaneous and remote access to scholars, even on mobile devices, following either single-user or collaborative schemes. Furthermore, the ability to incorporate digital copies in simulations can widen the study possibilities and the awareness of the possible integration of real objects in their actual context. • Effective cultural dissemination and education. Virtual exhibition and museum technologies, particularly those on the Web, constitute one of the most effective ways to disseminate, promote, and educate about cultural heritage. This approach offers to a wide audience the opportunity to familiarize with the richness of cultural heritage while removing certain barriers like long distances to the actual museums, spatiotemporal limitations, “do not touch” signs, linear narratives and predetermined guides, poor textual-only descriptions, and more. On the contrary, they may support rich multimedia interactive presentations, nonlinear and personalized narratives, and fully featured object examination capabilities at home or at school. In addition, virtual reality (VR), augmented reality (AR), and

47

3D Digitization of Tangible Heritage

1367

gamification approaches can significantly enrich the presentation and support the imprinting of long-lasting experiences. • Effective diplomacy, tourism, and economical development. 3D content and application, particularly on the Web, can be a very important toolkit for cultural diplomacy, tourism, and the support and strengthening of regional economies in an era, in which the diversity and uniqueness plays a significant role in the globalized world.

47.3

3D Digitization Methods

Today there are a number of 3D digitization systems based on a variety of different methods and approaches. Different technology solutions have been proposed, taking into account features such as high reflectance surfaces, color diversity, and morphological-geometric complexity. The accuracy of digitization, the ease and speed of use, and the range of materials that can be captured are among the main motivations that guide and sustain the development in the domain. A large amount of the commercial systems are based on laser triangulation, mainly because this approach achieves geometric results of high accuracy and resolution. Other popular methods include pattern projection and detection, photographic and photogrammetric methods, but also special methods such as those that use touch-sensitive robotic arms, which continue to serve the needs of special applications. The method chosen at each time determines, in practice, both the scan field and the expected results. A basic objective of any 3D digitization method is to measure the distance to the surface of the measured object, a quantity typically called depth. The motivation for various depth measurement methods was given by the depth perception capabilities and limitations recognized in humans, like in the study of [1]. On one hand, the multitude of different available methods reveals a strong research interest in this subject. On the other hand, those different solutions reveal the complexity of the problem. Geometrical principles, computer vision, filtering, and optimization algorithms have made the main contributions in the field. The 3D digitization methods that can be found in the scientific literature and already have practical applications can be categorized as follows: • Measurement techniques using lasers – Laser triangulation – Time of flight scanning • Shape/structure from X – Shape from structured light – Shape from silhouette – Shape from stereo – Shape from texture – Shape from shadow – Shape from shading – Shape from photometry/photometric stereo

1368

G. Pavlidis and A. Koutsoudis

– Structure from motion – Depth from focus/defocus • Traditional approaches – Traditional photogrammetry – Topography – Empirical techniques • Special techniques – Digital holography – Atomic force microscopy, stereo scanning electron microscopy, confocal microscopy, white light scanning interferometry – X-ray tomography The following paragraphs provide descriptions of the most widely used optical methods (laser-based and shape-from-X) and information regarding their advantages, disadvantages, application, and limitations. The chapter concludes with a reference to the influence of tangible heritage object characteristics, particularly the object size, in the selection of the most appropriate digitization method.

47.3.1 Short-Range Laser Beam Triangulation Short-range laser beam triangulation or simply laser triangulation methods are typically simple systems consisting of a laser source, emitting usually in the visible spectrum, and an optical sensor (a photographic camera) that is used to detect the laser light as it is reflected by the surface of the measured object. The principle of such a system is graphically depicted in Fig. 47.1. The location and orientation of the laser source and the detector are known; thus, a, θ, ψ are known quantities. As shown in the figure, the measurement process naturally forms triangles with a known base (the baseline) and the two of the three angles. The geometric characteristics of the triangles have been the basis for many measurement techniques, from basic geodesic measurements in ancient Greece to today’s 3D scanners. The most fundamental geometric laws applicable to 3D digitization are the laws of sines and cosines, which provide an immediate solution to the problem of triangulation. As shown in

Fig. 47.1 Principle of operation of short-range laser beam triangulation systems

47

3D Digitization of Tangible Heritage

1369

Fig. 47.1, calculating the distance of a surface point from the 3D scanner b is theoretically easy based on the known angles θ, ψ and the baseline a, using the law of sines, h i a b c ¼ ¼ ) sin ϕ sin ψ sin θ a b ) ¼ ) sin ðπ  θ  ψ Þ sin ψ sin ψ )b¼a sin ðπ  θ  ψ Þ

ð47:1Þ

The law of cosines for this case is expressed in any of the forms, a2 ¼ b2 þ c2  2bc cos ϕ b2 ¼ a2 þ c2 ¼ 2ac cos ψ 2

2

ð47:2Þ

2

c ¼ a þ b  2ab cos θ Figure 47.2 depicts how a distance is transformed into a displacement of the detected laser beam (at p1) from the optical center (c) on the 2D surface of an optical detector (typically a photographic camera sensor), resulting in an accurate distance (or depth) calculation. As shown in the figure, the location at which the laser pulse is focused on the image plane changes with a change in the surface point location (see, e.g., the fainted lines and points in Fig. 47.2 resulting in a new image point p2). Apparently, the accuracy of the depth measurement is directly related to the accuracy in locating the focused beam on the image plane. Laser sources are used in 3D scanning due to their special characteristics, basically their small dispersion, their high and concentrated power that is maintained over long distances, and their strict monochromatic emission. These characteristics make lasers ideal for contactless 3D digitization. Although many laser scanning systems are based on point sources, there is a variety of different patterns. Typically,

Fig. 47.2 Laser beam triangulation expressed as a displacement on the detector image plane

1370

G. Pavlidis and A. Koutsoudis

laser-based triangulation systems are most useful for distances of less than 10 m, due to practical limitations including (a) the gradual dispersion of the beam with distance and (b) the limited resolution provided by the imaging sensor. As expected, resolution and accuracy decreases with the distance. The resolution achievable in laser triangulation systems is theoretically a function of the magnitude of the angular changes that the system can detect in relation to the laser beam radius that is projected on different points on a surface, and the size of the radius of the reflected beam. In tangible heritage applications where high accuracy and resolution results are required, the best focusing ability of the beam should be targeted. In addition, accuracy and resolution are limited by diffractions and refractions. In fact, the thinner the laser beam, the greater the deviation created by these deviations, with diffraction playing a major limiting role on the possible radius of the beam. An in-depth analysis of the characteristics of light beam propagation in such cases can be found in [2]. It should be noted that in systems where depth is calculated from a single laser pulse, measurement errors occur due to the degree of reflection of the beam on the surface of the object, which is related to the nature of the surface itself [2]. Nevertheless, in any case, the laser beam is expected to exhibit a Gaussian profile on the surface of the measured object, and this phenomenon is also known as speckle effect. Responsible for this phenomenon is the surface itself, predominantly, its roughness. As a result speckle noise is inevitably transferred to the measurements. In addition, as laser triangulation systems rely on a digital imaging system to attain continuous world coordinates, a quantization process occurs that further limits the resolution. Speckle and quantization noise are thus imprinted in the measurements, and spatial filtering is typically applied to reduce these types of errors. However, the filtering approaches are most effective when simple planner surfaces are considered [2] and, thus, once again becomes apparent that the surface features affect the performance of a laser triangulation system [3, 4]. Surface characteristics are of paramount importance and have been thoroughly studied in the context of laser triangulation. It has been established that Lambertian (matte) surfaces are the most “friendly” surfaces for laser triangulation. Specular surfaces tend to deflect the beam to undesired angles (non-detectable by the system), whereas translucent surfaces tend to let light penetrate into the object’s substrates. In the latter case, light diffuses into substrates and creates internal dispersion, resulting in multiple unwanted reflections [5]. In addition to the measurement and quantization noise, one should not forget the electronic noise inherent in all electronic systems. Several variations of the basic point-based triangulation system have been proposed in the past, including various laser patterns, the usage of rotating or static mirrors to precisely guide the laser beams, or even the usage of complex systems of multiple mirrors, the usage of more than one optical sensors to improve the efficiency of the scanning system (not the accuracy though). Rioux [6] presented an innovative approach in a synchronized rotary system that can scan using very long focal length lenses with very small triangulation angles, without compromising the measurement accuracy. Small angles reduce the probability of occlusions. In cultural heritage applications, apart from the geometric information it is essential to capture surface color information (texture) for a complete digital recording. Laser triangulation systems are typically adapted to solve the geometric and not

47

3D Digitization of Tangible Heritage

1371

the spectral problems. Thus, in many systems, a dedicated digital camera system is also included to capture the object surface color. Apparently, in these cases, the two systems have to be synchronized so that the recorded texture can be assigned to the recorded geometry, which points to careful calibration processes [7].

47.3.2 Shape from Silhouette The method now called Shape from silhouette first appeared before the digital age, in the 1860s, when François Willéme presented a method called photo sculpture [8]. In photo sculpture, 24 photo shooting locations were used to obtain a corresponding set of silhouettes of an object (from 24 different angles). Development of the silhouettes was done on photographic plates, which were then projected with the help of an image projector named magic lantern. The silhouettes were then transferred to clay using a pantograph. Commercial applications of photo sculpture were immediately developed, and in large capitals, such as Paris, London, and New York, specialized photo studios were established that offered them as services. The basic shortcomings were that the process required a high degree of human intervention, the cost of necessary equipment (cameras and image projectors) was high, and there was a requirement for an operator with expertise. Ultimately, the lifetime of photo sculpture was short (1863 to 1867) mainly due to its high cost. With the advent of the digital age, photo sculpture was, in a way, transformed into a digital method named Shape of silhouette. During the first years of the development of 3D digitization methods, Shape from silhouette was a very popular technique for movable objects. Its initial implementation was presented in the context of a doctoral thesis [9]. Later, another approach presented the creation of 3D models from multiple angles using volumetric data [10]. The methods were based on extracting silhouettes of objects from series of photographs and gradually creating the 3D digital replica through the analysis of the silhouettes. Essentially the method consists of a camera fixed on a tripod opposite to the measured object that is placed on a special rotating base (usually called “turntable”) with a precisely controlled rotation step. The background behind the object is monochrome, so that the silhouette can be easily separated in each photo [11]. Figure 47.3 graphically shows the side view of the typical Shape from silhouette setup. According to studies, distance D should average to about 120 cm, while d should be about 3 mm (ideally zero) [12]. The total number of silhouettes plays an important role in the quality of a reconstructed object, as some of the details of the object are only visible from specific angles. As the method is based on silhouettes, it should be noted that the resulting 3D model will not include morphological features that are not visible on them. Apparently, the resulting 3D model is an approximation of the original object rather than an exact digital replica. A crucial step before capturing photos is the calibration of the system, which provides the external camera parameters (location, orientation), as well as the internal camera parameters (focal length, displacement of the optical center). The calibration process uses a standard test object with a specific pattern, usually a “chessboard,” with known geometric features.

1372

G. Pavlidis and A. Koutsoudis

Fig. 47.3 Side view of a typical Shape from silhouette setup

The silhouette extraction in the photos relies purely on well-established robust computer vision (mainly image processing) techniques. A simplified approach is based on taking photos of the object against a brightly lit background so that the object appears virtually black in front of a virtually white background. Using a reference photo for the background without the object makes it easy for the computer vision algorithms to be able to easily separate the object from the background pixels with simple thresholding approaches. In more complex approaches this thresholding might need to be performed separately for each of the color channels of the photos. As defined in [13], a single silhouette indicates the 2D image area, in which all the visible points belonging to the measured object are included. Considering that there is no further knowledge about the position of the measured object in space, then all the information given by a silhouette Si, in a set of i ¼ 1, . . ., n silhouettes, is flat over an area of space Ci. This is retrieved through back projection of Si from the corresponding viewpoint Vi. The n silhouettes limit the object within a volume Rn ¼ \ni¼1 Ci . Apart from the simple pixel-level silhouette extraction approach, there is also a region-based silhouette extraction approach, which can improve the accuracy of the results, and is based on image segmentation into regions and a subsequent regionbased image thresholding. This approach was developed to tackle cases, in which object surface colors match the background color, like when the surface is shiny and some regions may reflect a high intensity light back to the camera, rendering pixellevel thresholding approaches incapable of recognizing the phenomenon, thus creating holes in the silhouette. The final step in the process is the recovery of 3D geometry from the multiple silhouettes. Two approaches used for this task are: • The virtual “carving” of the volume of the object by gradually imposing the volume limitations that every silhouette introduces [13]. In this approach the object is represented by a geometric entity known as a visual hull. Each silhouette is projected onto and inside the visual hull. This projection is either perspective or

47

3D Digitization of Tangible Heritage

1373

parallel. The 3D digital model is created by the sections of the volumes that result gradually from the projection of each silhouette. This approach is called space carving [14] and consists usually of two phases: (a) rejection of non-object parts and (b) reconstruction and organization of vertices into meshes (surfaces) [14–20]. • The gradual creation of the 3D model mesh with a virtual “carving” of an initial mesh from a silhouette, which is gradually improved by the subsequent silhouettes being examined. In this approach silhouettes are projected onto a mesh with simultaneous application of direct section. The common space defined by any two views is calculated by the logical AND operation that corresponds to the intersection of the two 3D projections (3D Boolean intersection). Silhouette after silhouette the mesh is carved to result the final 3D reconstruction of the measured object. After the recovery of the 3D geometry, another process initiates to retrieve the texture information (color). This is done by repeating the photo capturing process for the same locations but with different lighting conditions. In this step the object is properly lit with sufficient ambient light (to limit shadows and high-intensity reflections). The mapping of color onto the 3D mesh is done on the basis of matching each mesh triangle normal vector with the visual axis of the photos, thus using the photo that is most aligned with each triangle in the mesh [13]. The fact that various photos may be used for the coloring of adjacent triangles in the 3D mesh is the source of color distortions and chromatic discontinuities. In addition, no matter how carefully the lighting of the object is selected, there will still be unwanted shadows casted by the complex object geometry, which will also introduce color distortions. Special color blending techniques have been developed to ensure smooth color transitions between photos and particularly at the photos’ borders when those colors are mapped onto adjacent triangles in the 3D mesh (e.g., [14]). Shape from silhouette is very likely to produce satisfactory result for translucent objects (like glass) or for objects with high reflectivity (like jewelry), if carefully selected lighting conditions are imposed. In some cases, manual intervention may also be necessary to control and correct parts in the silhouette images. The main disadvantages of Shape of silhouette are (a) the fact that the final 3D digital replica only carries the information contained in the silhouettes. Cavities and holes on the surface that are not visible in silhouettes do not contribute to the creation of the 3D model, and inevitably the final 3D model is an approximation of the original object [13]; (b) the number of silhouettes used to produce the 3D replica plays an important role in the quality of the result. In theory, one might argue that as the number of silhouettes increases, the smaller the geometric error becomes. It should be stressed here that in tangible heritage digitization applications where geometric accuracy and resolution are crucial, Shape from silhouette should be very carefully planned and applied, due to the inherent limitations of the method discussed above. Multiple viewing locations and carefully planned rotation steps are typically needed to capture the geometric complexity of cultural objects. Nevertheless, if the specific application targets digital or online exhibition goals, then this method is highly productive and automated and could be significantly advantageous.

1374

G. Pavlidis and A. Koutsoudis

47.3.3 Shape from Stereo Shape from stereo is a method that attempts to extract structure information from a single pair of photos, with many applications in the field of computer vision and robotics. Shape from stereo implements an analogy of biological visual systems and particularly binocular vision or stereopsis; the method uses a pair of photos (a stereo pair) that show the same scene from positions that correspond to the positions of two eyes. In a stereo pair, the contents of the two images largely overlap, while the displacement of one image relative to the other is on a single axis that coincides with the axis of the baseline (the line joining the two eyes). This displacement is usually expressed as disparity between the images in the stereo pair and directly connects with the various distances (depth) in the viewed scenes, which is another way of seeing triangulation. Figure 47.4 shows a simplified representation of a binocular visual system from which it is easy to derive the disparity that arises from the two views for a single point in space. From the similarity of the triangles in the setup, the disparity relates directly to depth as follows:

Fig. 47.4 Simplified binocular vision system

47

3D Digitization of Tangible Heritage

EF BE ¼ BD AD GH CG ¼ CD AD BE ¼ CG

1375

9 > > > > > > > = > > > > > > > ;

) d ¼ EF þ GH )

d ¼ EF þ GH BD CD d ¼ BE þ BE ) AD AD   BD þ CD d ¼ BE ) AD bf d¼ z

ð47:3Þ

where b ¼ BC is the baseline of the two-camera system, f ¼ BE ¼ CG the focal length of the cameras lenses (same for both), and z ¼ AD the distance (depth) of the measured point A from the baseline. Apparently, this simplified presentation can be a good approximation of reality only if the cameras are exactly the same, and their relative positions and orientations are strictly parallel facing toward the scene with the same angle. In reality these conditions are not met, and thus the cameras should undergo a calibration so that the system, in a way, approximates the simplified case shown. Using (47.3) depth can be recovered from disparity as z¼

bf d

ð47:4Þ

given a known baseline and a focal length. For all practical purposes, the numerator can be disregarded since it is a constant, and thus depth can be considered to be simply the inverse of disparity. It is clear that in order for depth to be recovered from pairs of photos, point correspondences should be resolved first. That is to identify the points on object surfaces in both photos. This is the well-established point correspondence problem and is pervasive in a large number of computer or robotic vision applications. The solution to this problem can be significantly sped up by imposing epipolar constraints. The epipolar constraints arise from the topology of the cameras+object system and restrict the regions in photos in which to search for point correspondences (a point in one camera is restricted onto a line in the other camera). In any case, before searching for point correspondences, there is still the need to either rectify the photos [21] (virtually, rectification of the photos means to distort the photos so that they correspond to cameras that are of parameters equivalent to the simplified case of Fig. 47.4. A graphical representation of the photo rectification is shown in Fig. 47.5.) or use the intrinsic and extrinsic camera parameters to correctly apply the epipolar constraints.

1376

G. Pavlidis and A. Koutsoudis

Fig. 47.5 Rectification of photos for stereo vision

Apart from the epipolar constraints, a number of other constraints may be used to improve the depth estimates, including the similarity and uniqueness constraint, as well as the continuity and the ordering constraint. The similarity constraint is straightforward and relates to correlation. The uniqueness constraint imposes that a point in one image should correspond to one and only point in the other image. The continuity constraint imposes that disparity should be piecewise smooth. The ordering constraint imposes a particular ordering of the identified points. It should be emphasized that the solution to the point correspondence problem is not as simple as expected, since images are usually corrupted by noise, occlusions, and lighting variations; thus the point localization is essentially guided by error minimization approaches (typically, there is no perfect match). Nevertheless, a typical depth recovery process is done by a series of algorithms, as shown in Fig. 47.6. The end result is a depth map. Each pixel of the depth map describes a distance of a point on the measured object’s surface from the cameras. Shape from stereo is a very interesting method for the research community, which is revealed by the large number of works that have been published regarding point correspondences [22] and the applications in computer and robotic vision.

47

3D Digitization of Tangible Heritage

1377

Fig. 47.6 Generalized depth map generation approach in Shape from stereo

Correspondence algorithms can be generally categorized into feature-based and correlation-based. In feature-based matching, the relevant algorithms undertake to locate all those distinct features (angles, lines, curves) on the surfaces of the objects that are visible in both photographs. Although the approach leads to robust algorithms and implementations, depth recovery is only possible for the identified features, thus leading in general to a sparse depth map. In correlation-based matching, image blocks of fixed size are being compared, and their similarity is assessed in terms of their correlation, resulting, typically, in dense depth maps. The main advantage of Shape from stereo is that the method is easy-to-use and relatively automated, without any need to use particularly specialized equipment or conditions for data collection. On the other hand, the main disadvantages of the method are (a) depth is calculated only on point correspondences, (b) the resulting depth maps contain data that include intense noise due to errors in point localization, and (c) a complete 3D digital replica would require a large number of stereo pairs to cover the entire surface, thus a number of processing steps like alignment and consolidation of partial scans. Shape from stereo has a number of limitations that should be taken into account, since they hamper the process of depth recovery. These limitations include matching issues in areas around the boundaries of the objects, areas on the surface of an object with strong light diffusion phenomena (non-Lambertian surfaces), occlusions on one of the cameras, and the effects of the perspective projection on the viewed scenes [22, 23].

47.3.4 Shape from Structured Light Shape from structured light is a method in which depth recovery is based on the projection of a light pattern on the surface of the measured object. Estimates of the deformations of that pattern on the surface lead to the recovery of the depth (distance from the object). Basically, just like with laser scanning, depth is obtained through triangulation. A fundamental difference between the two approaches is that Shape from structured light makes use of typical image projection systems (projectors) instead of laser sources. A simplified diagram showing the principle of operation in Shape from structured light is shown in Fig. 47.7. The light pattern consists of three vertical straight colored stripes (red, green, and blue, respectively) which are

1378

G. Pavlidis and A. Koutsoudis

Fig. 47.7 Principle of operation of Shape from structured light

projected onto a measured surface. The pattern is deformed due to the surface geometry and is reflected to a detector (typically a digital camera). The color coding of the stripes in the projected pattern makes it easier to recover the geometry. It should be noted that the encoding of a pattern need not be based on color but also on intensity, shape, and density. The triangulation is performed pixel-wise, on every pixel of the detected image. Apparently, in order to be able to apply the triangulation, triangles should be formed that connect each pixel in the detected image with a single point on the surface of the object and a single pixel on the projector, something that is not straightforward without a proper calibration prior to the application of the method. Practically though, the problem is addressed as a stereo vision problem, in which the second camera (of the virtual camera stereo pair) coincides with the projector; thus the estimates include the notion of disparity (the displacement of points on the image plane due to their different distance from the camera), transforming the triangulation into a disparity-depth estimation. The disparity for a point in space on the detector camera image plane is connected with the distance as defined in (47.3), in which the second camera has been replaced by a projector; thus b is the baseline connecting the projector with the camera, and f is the focal length of the camera, which should be the same for the projector in a rectified system. Given a reference point at distance zref which exhibits a disparity dref, then the distance at which another point lies can be computed using (47.3)

47

3D Digitization of Tangible Heritage

1379

bf bf  ) z zref bf bf ¼ d  d ref þ ) z zref 1 d  dref 1 ¼ ) þ z zref bf 1 z¼ d  dref 1 þ zref bf d  dref ¼

ð47:5Þ

In addition, the rate of change of depth (distance) with disparity, from (47.4), can be computed as @z bf z2 ¼ 2¼ @d bf d

ð47:6Þ

from which the discrete depth resolution that such a system may offer is connected with the disparity and thus the image resolution of the camera Δz ¼

z2 Δd bf

ð47:7Þ

which implies that an expected nominal resolution corresponds to a Δd ¼ 1 disparity resolution, although methods have been proposed that achieve sub-pixel resolution [22, 24]. Today, there are several structured light techniques, categorized according to the stripes’ encoding being followed, including wavelength, range, temporal, and spatial multiplexing [25]. Wavelength multiplexing is the already referenced color coding of the stripes. A special case of wavelength multiplexing is the method of direct coding [26], in which each pixel within a scan-line is associated with a specific color value and the disparity of the pixel can be estimated by matching image windows of a single pixel. Range multiplexing is the approach in which the range is characterized by multiple gray values. Temporal multiplexing is a popular technique, in which a set of different patterns are being projected one after the other. These patterns consist typically of vertical white and black stripes the spatial frequency of which changes with time (i.e., the stripes start as wide as possible and become thinner and thinner at every step). Spatial multiplexing is the most popular technique, usually referenced as active stereo vision or projected texture stereo, in which De Bruijn patterns are typically used, which can be tiled according to the needs of the application. Several approaches can be found in the literature using variations in the encoding and setup. In general though, in systems that implement the Shape from structured light method, the creation of a 3D digital replica occurs through data processing including alignment, partial scans consolidation, and texture blending [27–30]. As with other methods, the texture (color) information is derived from a set of photos

1380

G. Pavlidis and A. Koutsoudis

collected from the same viewing angles with the pattern projections, as the measured object is photographed both under the projection of the patterns and under natural illumination to capture the geometry and the spectral information. The 3D digitization accuracy of such systems can be affected by (a) the quality of the optics of the projection system, (b) the focal length and the range of the projection system, (c) the sensor quality and the quality of the optics of the digital camera, as well as the sensor resolution and the fidelity of color recording, (d) the permissible viewing angles of the measured object (to avoid shadows caused by the system itself), and (e) the surfaces of the measured object and its characteristics [27, 28].

47.3.5 Shape from Shading Shading plays a very important role in the perception of depth. It is one of the depth cues exploited by the visual system in order to deduce distances purely from differences in image intensity, which the visual system attributes to the surface geometry. Apparently, color is irrelevant in this case and only gray-level images are considered. Intuitively, as light hits the surface of an object and is reflected to an imaging system that detects luminance, it creates images of various shades of gray that depend on the geometry of the surface, following the changes in the emergent angle of light. The image intensity can be related to the light intensity in the scene by I ðX, Y Þ ¼ κLðx, y, zÞΦðn, s, vÞ

ð47:8Þ

where I is the image intensity on a pixel at (X, Y ), L is the light intensity on a 3D point at (x, y, z), Φ the reflecting properties of the surface as a function of the surface normal n, the light source direction s and the viewing direction v, which is typically defined for a Lambertian surface as Φ(n, s, v) ¼ ρ cos i (ρ being the albedo and i the incident angle). Assuming a fixed setup in which the measured object the detector and the light source are in known fixed locations, the resulting image intensity is a function of the surface normal n. Typically, the surface normal is represented in p  q gradient space, which represents the rate of change of z coordinate (depth) with  @z @z respect to x for p and y for q (as z ¼ f(x, y) and the surface gradient is @x , @y , 1 ,

@z @z and g ¼ @y ). This forms what is known as the reflectance map of thus defining p ¼ @x an imaged scene R( p, q). Thus, formally stated, the aim in Shape from shading is to recover depth from the orientation of a surface element on a measured object given the reflectance map of the object’s surface. Shape from shading is a long studied problem and originates in the work of Horn in 1975 [31], in which it was approached as the solution to the brightness equation, the nonlinear first-order partial differential equation that connects image intensity with the scene reflectance map, explained above and usually expressed in the form I(x1, x2) ¼ R(n(x1, x2)), with (x1, x2) the image space coordinates of a point. Shape from shading has been addressed by many works and today is known that the Shape

47

3D Digitization of Tangible Heritage

1381

from shading problem is an ill-posed problem without a unique solution and a number of publications focused on this issue and tried to provide workarounds using assumptions (like that the object surface is Lambertian) [32–38]. A mathematical formulation and a description of the main difficulties of the problem and some interesting theoretical results backed with realistic examples and numerical methods can be found in [39]. Shape from shading is applied using four approaches, differentiated by the algorithmic method of depth recovery from shading: • Minimization Techniques: minimization techniques result in solving the problem by using an energy function. • Transmission Techniques: transmission techniques study the transmission of shape information from a set of feature points located on the surface of the object. • Area Techniques: area techniques retrieve depth through assumptions about the surface type. • Linear Techniques: linear techniques calculate the solution of the problem based on mathematical mapping. Apparently, regardless any assumptions and workarounds, Shape from shading can only work if the homogeneity assumptions hold, that is, the setup is fixed, invariable during the process and of known parameters. In addition, a single view can only recover the depth from that specific viewpoint, and thus several photos of the object should be acquired from different viewing angles to cover the entire surface. In almost all cases, it is assumed that a light source follows a particular path to create a certain type of shading on the surface of a measured object. A major disadvantage of the method is the inability to extract depth from dark areas as they do not provide usable intensity information. The mathematical models for the reflected light are usually simplistic, while more complex models have already been proposed. Combinations of this technique with other techniques have been proposed to improve the results, like the combination with Shape from stereo [40] or with Shape from shading, which improves results in dark areas.

47.3.6 Shape from Photometry Shape from photometry or photometric stereo is the generalization of Shape from shading. In Shape from photometry the measured object is imaged under varying (but controlled) lighting conditions. The method uses a sequence of photographs showing an object from a fixed viewing angle but under varying lighting conditions (typically, a light source moving in a predetermined path). As in Shape from shading, the challenge is to recognize the surface orientation (or the surface normals) by the image intensity. The benefit of Shape from photometry is that the different light source locations create constraints for the possible surface orientations thus resulting in more robust depth recovery. In this method depth is recovered through intensity images. The surface normal and vectors involved in the method are shown in the

1382

G. Pavlidis and A. Koutsoudis

Fig. 47.8 Shape from photometry principle and the involved normals

simplified snapshot of the method depicted in Fig. 47.8, where n is the unknown object surface normal, ni is the image plane normal at the detected point, and ns is the light source vector. Shape from photometry is based on core notions on radiometry, and the interested reader should refer to the relevant bibliography for more in-depth intuition. Among the most basic concepts include definitions of the solid angle, the surface normal (a vector perpendicular to a surface), the radiance of a light source (luminous power per unit solid angle and projected surface), the irradiance of a surface (intensity received – a ratio of the luminous power per unit surface), the bidirectional reflectance distribution function (the relation between incident and reflected light), the Albedo (or diffuse reflectance), and the image intensity (see, e.g., [41, 42]). Photometric stereo has been thoroughly studied since its introduction by Woodham in 1980 [43]. Woodham defined the reflectance map as an entity related to the bidirectional reflectance distribution function (BRDF), defined in a US national stan@z @z dard [44]. He used the p and q parameters (p ¼ @x and q ¼ @y ) in the notation of the surface normal vector ( p, q, 1) thus forming a two-dimensional gradient space of all such points ( p, q). The reflectance map was defined as R( p, q) the map that determines image intensity as a function of the gradient coordinates. Woodham solved the problem for the case of orthographic image projection and a fixed distant light source. For a Lambertian surface with a reflectance factor ρ the reflectance map was defined as ρð1 þ pps þ qqs Þ Rðp, qÞ ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffip ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 þ p2 þ q2 1 þ p2s þ q2s

ð47:9Þ

where ( ps, qs, 1) the vector that points to the direction of the light source. This led to the description of image formation as a simple equation relating scene reflectance with image intensity I ðx, yÞ ¼ Rðp, qÞ

ð47:10Þ

47

3D Digitization of Tangible Heritage

1383

Based on this, Woodham proposed the photometric stereo technique in which he used the three source photometry approach (Horn B (1978) Three source photometry. Personal communication) to show that with three images (three different light source locations) determination of the reflectance factor ρ and the surface orientation is possible and unique. In the original notation, the image formation process is described as I ¼ ρNn ð47:11Þ h i⊤ ð1Þ ð2Þ ð3Þ where I ¼ I ðx,yÞ , I ðx,yÞ , I ðx,yÞ is the vector of the intensity values at image point (x, y) of the three images, ρ is the surface reflectance factor, N ¼ [n1, n2, n3]⊤ is the matrix that consists of the three light source normal vectors, and n is the surface normal that corresponds to image point (x, y). Reversing to solve for n, n¼

N1 I ρ

ð47:12Þ

In addition, as the equation there is an apparent need for N1 to exist, this can happen if and only if the three light source vectors do not lie in a plane, thus resulting in ρ ¼ |N1 I], overall producing n¼

N1 I j N1 I j

ð47:13Þ

Apparently, a final step in the process includes the transformation from surface normals to shape (like solving shape from normals). This can be approached from various angles depending on the projection model adopted (orthographic, perspective, etc.) and the lens optics (i.e., the focal length) and includes the solution of partial differential equations that simplify to difference equations in the discrete image domain. These solutions are accurate up to a scale, which means that the reconstruction is in an unknown scale. Woodham presented some illustrative examples and summarized the advantages over the typical stereo-based approaches in that the difficulty to identify image correspondences is removed, the effect of surface orientation on image intensity can be removed, and the measured object shape surface is determined by its surface orientation and not the distance (depth). In addition, Woodham already proposed a number of alternative implementations including (a) a moving light source, (b) multiple calibrated light sources, (c) rotation of the object and the camera, and (d) multiple colored light sources. Soon after the first applications in Shape from photometry, it was realized that the assumptions being made can only support the method for Lambertian surfaces; thus, variations have appeared to tackle other types of reflection models. Examples include color images and different color sources that showed how the accuracy and robustness are improved and materials that are composite, dielectric, and colored metal could be handled [45]. An approach that uses reference geometric solids that are photographed alongside the measured object was proposed in [46]. According to

1384

G. Pavlidis and A. Koutsoudis

the authors, this approach made it possible to create high-quality 3D reconstructions that were comparable (even better) than those created by laser triangulation methods. A practical description of the implementation of photometric stereo along with code can be found in [47]. An interesting application has been reported in 2012 in which the researchers aimed at determining the shape of a satellite based on photometry from four telescope locations [48]. Recently, a portable low-cost version has been developed for smart mobile devices, named “Mobile Shape-from-Shifting,” that is applicable to tangible heritage and is able to eliminate specular reflectance errors using only simple polarizer filters on mobile devices. It should be emphasized though that this method is a hybrid as it adopted both geometric and photometric shape recovery [49]. Shape from photometry can be used for practical applications in tangible cultural heritage, since it is a non-contact high-quality measurement method, but requires special lighting and controlled environment which might be infeasible in some cases for on-site work.

47.3.7 Shape from Texture Texture is an important source of information that aids in perceiving the morphology of the surface of an object. Shape from texture is a technique based on the assumption that there is knowledge about the structure and morphology of the surface texture of the measured object; thus deformations of this structure can easily be used as a cue for the surface geometry. Particularly when the surface is composed of tiled elementary texture patches, usually termed texels ([tex]ture [el]ements), the process can be remarkably successful [50–54]. In order for Shape from texture to be efficient, robust recognition of the texture “structure” is required, which includes the identification of regularity, periodicity, collinearity, parallelism, orthogonality, and symmetry [54]. In summary, all the assumptions adopted are grouped into homogeneity (uniform distribution of texture over the surface) and isotropy (the texture consists of line segments with no preferred orientations) [52]. Shape from texture has a long history in research, and among the pioneering works was the work of Gibson in 1950 [55], in which the concept of the texels was introduced in a new theory of human visual perception of surface orientation. Shape from texture has no direct practical application in tangible cultural heritage, as it is very unlikely to have cultural objects with a single repeating pattern on their surface. Practical applications that appear to be related to this method can be found in the reconstruction of fabric surfaces, human skin, and cork [53].

47.3.8 Depth from Focus/Defocus Of particular interest is a method that can recover the scene depth from a set of narrow-depth-of-field images using the degree of focus (or defocus) as a distance

47

3D Digitization of Tangible Heritage

1385

cue. The method differentiates significantly from photometric and geometric methods, as it does not rely on point correspondences or spectral responses of surfaces. The basic principle of operation of Depth (or Shape) from focus/defocus is easier to understand by adopting the thin lens model in geometrical optics, as shown in Fig. 47.9. In this representation, A is the diameter of the lens aperture, f is the focal length of the lens, uo is the distance between the lens and the object point P that is imaged, ui is the distance between the lens and the location of perfect focus Q, v is the location of the image plane relative to the lens, and c is the circle of confusion or blur circle on the image plane that represents the disc that corresponds to a single object point. In this representation, 1 1 1 þ ¼ uo ui f

ð47:14Þ

By exploiting the properties of similar triangles in Fig. 47.9, c ui  v ¼ A ui

ð47:15Þ

Apparently, the bigger the disc c the worst the blurriness of the imaged point. In the thin lens model the aperture and the focal length are connected by A¼

f N

ð47:16Þ

where N is the f-number of the lens diaphragm. Putting together (47.14), (47.15), and (47.16) results

Fig. 47.9 The thin lens model of geometrical optics and the circle of confusion

1386

G. Pavlidis and A. Koutsoudis

9 f ui  v > uo f > v N ui = f uo  f )c¼ ) uo f N u f > ; ui ¼ o > uo  f uo  f f uo þ fv þ uo v c¼ N uo c¼

ð47:17Þ

Finally, solving for the distance to the object point (or the depth), uo ¼

fv cN þ v  f

ð47:18Þ

Special attention should be paid to having all quantities measured using the same units; c could be expressed in pixels, although distances could be in mm; thus it is highly likely that unit transformations should take place. Depth from focus is considered to have been originally proposed in 1985 independently by two researchers Pentland [56] (later also in [57]) and Grossmann [58]; Grossmann submitted his work in 1995 but it was finally published 2 years later. Pentland [56, 57] was mainly concerned with the development of the mathematical formalism for the extraction of depth information from focal gradients, but also presented experimental results. He relied on the thin lens model of geometrical optics to derive the basic formula that connects the degree of focus with the distance. He formulated (47.18). Pentland offered two approaches to estimate the amount of blur (the size of c), one based on sharp edges and another based on using images of the same scene with different lens apertures. Grossmann [58] on the other hand was less concerned with the definition of a mathematical formalism. Hismain concern was the development of a simplified, practically useful implementation approach. This method was based on (a) locating image primitives, like edges, (b) evaluating a blur measure for each primitive and (c) converting the blur measure to relative or absolute depth. The method was also based on the thin lens model of geometrical optics. Grossmann created an indicative image on which he presented his definition of the blur measure w as the number of pixels on image edges and provided his basic formulas of computation as (in his own notation) uo ¼ v  κ  c

ð47:19Þ

where uo is the absolute depth, κ is a camera setup constant, κ  c is the relative depth, and c is the blur parameter defined by w2 ¼ w2i þ c2

ð47:20Þ

in which wi is the smallest value of w observed, assuming a convolution of two Gaussians. Grossmann resolved the -ambiguity of (47.19) “by hand,” and proposed a calibration-based definition of the constant κ.

47

3D Digitization of Tangible Heritage

1387

About the same time, Subbarao proposed to change the intrinsic camera parameters as an alternative to the previous approached for depth recovery [59]. Up to this point, the method seemed to basically be “Shape from defocus.” In 1988 Darrell and Wohn presented the, most likely, first method that relies on focus (and not defocus) [60], which used Laplacian and Gaussian pyramids to calculate depth in a set of images with varying focal distance. Ohta et al. [61] presented the use of a set of images of varying focal distances in order to reconstruct a high-quality all-focused image, but they did not suggest the extraction of geometry for such images. In addition, Kaneda et al. [62] applied the same approach on sets of optical microscope images to recover depth and create pan-focused stereoscopic representations. In 1989, Nayar published a technical report entitled “Shape from focus” that has widely been used and cited, which offered two more practical implementation approaches [63]. This was another multi-image varying focus distance approach, in which the principle of operation is graphically depicted in Fig. 47.10. In all those cases, which rely on geometrical optics, the measured object is placed on a planar base that moves toward (or away from) the camera, having the initial (reference) and final distances from the camera known, along with the camera depth of field, the focus distance, and the movement step. Since, in this configuration, the captured photos present only some of the surface in perfect focus, the depth can be determined for every captured pixel, and thus from the reference distance d and the focus distance df the actual object geometry ds can be determined. In each photo only a few sharply focused pixels can be used to extract depth; thus the process is repeated for a number of photos taken at different focus distances. In this chapter, a new focus estimate was defined in order to assess the sharpness of image regions, and two implementation approaches were presented, one based on simply assigning

Fig. 47.10 Principle of operation of Depth from focus/defocus

1388

G. Pavlidis and A. Koutsoudis

depth to the distance that maximizes the focus measure and another that uses Gaussian interpolation to improve the depth estimate. Nakagawa and Nayar filed a patent that was published as US5151609A [64] in 1992 after being assigned to Hitachi in 1990. Ens and Lawrence in 1993, following the line of research on Depth from defocus, presented an in-depth investigation of the limitations of methods proposed up to early 1990s and proposed a matrix-based method that seeks to deconvolve the defocus operator from the images [65]. In this method, two images of the same scene are acquired with different defocus operators (i.e., aperture setting or focus distance) h1(x, y) and h2(x, y) (the latter being the operator that corresponds to the strongest defocus). The researchers transformed the problem of Depth from defocus into a problem of finding a third defocus operator (or convolution ratio) h3(x, y) that is a unique indicator of depth, such that, h1 ðx, yÞ  h3 ðx, yÞ ¼ h2 ðx, yÞ

ð47:21Þ

The researchers proposed three techniques to recover h3(x, y). In their general solution that is applicable to noisy image data, for any image regions in the pair of acquired images I1(x, y) and I2(x, y) the convolution ratio is approximated by a least squares minimization, min

XNk XNK h x¼0

y¼0

 I 1 ðx, yÞ  h^3 x, yÞ  I 2 ðx, yÞ2

ð47:22Þ

where h^3 ðx, yÞ is the approximation of the convolution ratio of size k  k, assuming an N  N image I1, and thus an (N  k + 1)  (N  k + 1) image I2. Xiong and Shafer in a Technical Report published in 1993 (also appeared in [66]) were among the first to distinguish and document the two flavors, Depth from focus and Depth from defocus. They provided a clear definition of both methods and proposed improvements for both cases. In 2000 Paolo Favaro also provided definitions, listed more than 70 works on the subject, and connected the problem with other well-known problems from the scientific literature, such as “blind deconvolution,” “source separation,” and “inverse scattering” [67]. In addition, there have been works that evaluated the performance of Depth from focus/defocus and compared it to other techniques like stereo and vergence (see, e.g., [67, 68]). In 1999 Chaudhuri and Rajagopalan devoted a book on passive methods for depth recovery, focusing mostly on Depth from focus/defocus [69]. Recently, fast implementations of Depth from focus/defocus have been targeted for improved performance on mobile applications (see, e.g., [70–72]). Last but not least, deep neural network architectures have been proposed as an end-to-end learning approach to the solution of the Depth from focus problem, and a DDFFNet has been created and was trained using light field camera (RGB-D) data [73]. DDFFNet provides depth map predictions in a fraction of a second and has reduced the depth error by more than 75% compared to previous methods. Shape from focus/defocus is a very interesting approach and can be easily applied in tangible cultural heritage applications. Today’s high-quality optics and digital

47

3D Digitization of Tangible Heritage

1389

cameras make it possible to create small depth of field photos, and a workflow similar to the one applied for focus stacking is easy to implement and deploy for artifacts. Nevertheless, the demand for very accurate depth estimates in cultural heritage does not seem to be able to be fulfilled by Shape from focus/defocus approaches as there is always an inherent error in estimating depth from defocus (or focus). In addition, Depth from focus/defocus reconstructs a depth map of an object from a single viewpoint, and thus multiple viewpoints should be used and the results should be consolidated into one final 3D objects.

47.3.9 Shape from Shadow Shape from shadow is a family of methods that use the shadows on objects’ surfaces as cues to recover the structure of the objects. These methods are either based on radiometric (in a broad sense) or on structured light approaches. The first class of Shape from shadow methods have long been investigated, as shadow is a pervasive depth cue in vision. Apparently, these approaches are efficient only in the assumption that the shadows’ begin and end points are accurately identified and self-shadows are being studied. The basic radiometric Shape from shadow idea is shown in Fig. 47.11, in which the setup is known and only selfshadows are being considered. Shafer and Kanade in 1983 presented a theory describing relationships among surface orientations in line drawings with shadows [74]. They analyzed several cases of shadow casting among objects under various lighting conditions and established fundamental constraints for the orientation of the surfaces. Hatzitheodorou and Fig. 47.11 Representation of the radiometric Shape from shadow

1390

G. Pavlidis and A. Koutsoudis

Kender focused on computing surface contours by using a directional light that casted sharp shadows [75]. They used multiple light source positions to be able to approximate the surface contours with interpolating splines (assuming surface contours are smooth). Later, in 1989, Raviv et al. presented a complete Shape from shadow approach, in which the measured object is placed on a reference plane with a camera above it and a light source moves in a circular motion above the object to enable capturing a set of images with varying shadows [76]. They too relied on shadow pair points to assess the height difference between these points. This method resulted a height field representation of the object surface. Langer et al. built on this work to compute holes beneath the recovered height field [77]. Yang in this PhD thesis was mainly concerned with improvements of the Hatzitheodorou and Kender method in cases in which error and inconsistencies appear [78]. A few years later, in 1998, Daum and Dudek developed a method that limits the possible height of each surface point and progressively improve the estimate by using information from more and more light source positions [79]. Savarese et al. in 2001 introduced “shadow carving” [80], a method that exploits the notion of volume carving using epipolar constraints to produce a “conservative” estimate of the measured object volume using shadows, and in the next year they presented a practical implementation of a system for shape capturing [81]. This research has been the main body of Savarese’s PhD thesis, and the complete theory and practice of the approach has been presented in 2007 [82]. Shape from shadow with shadow carving is robust with respect to errors in the detection of shadows, and it enables the complete object reconstruction, rather than the single-view 2.5D relief that was able with previous approaches. About the same time, in 2005, Yu and Chang focused on devising a new representation of the shadow constraints based on graph approaches [83]. Recently, researchers took Shape from shadow “outdoors” by defining the “episolar” constraint, the shadow pair constraint under sun light, and by assuming a known georeferenced camera location and accurate times-tamps (to account for known light source locations) [84]. In the second class of Shape from shadow methods, one may find a special case of structured light approach originally proposed by Bouguet and Perona in 1998 [85], which they reference as a weak structured light approach. In this approach the structure of a measured object’s surface is captured by locating the shadow of a known moving occluder (like a bar or simply a pencil) that is cast on the measured surface by a fixed light source. The concept is visualized in Fig. 47.12. As the occluder moves, the camera records the surface of the object with the shadow of the known occluder appearing in various locations on the surface. Deformations of the shadow due to the relief of the surface can lead to the depth calculation using triangulation. This method was proposed due to its simplicity and low cost. Figure 47.13 shows snapshots from an example application with an artifact and the recovered geometry. In this approach, the camera parameters, the light source location, and a reference plane are known a priori. In an interesting approach by Kawasaki and Furukawa [86], a hybrid of the two approaches has been presented, in which camera parameters, reference plains, and light source locations are unknown. The method works only with straight linear

47

3D Digitization of Tangible Heritage

1391

Fig. 47.12 Representation of Shape from shadow as a special case of structured light

objects or straight edges in the examined scene, which is a serious limitation, but the reported results are improved in terms of density and accuracy. Shape from shadow in its two forms can be easily applied in tangible cultural heritage digitization applications, but the results are expected to be of low accuracy, low resolution, or both. The structured-light form should be applied with caution as the width of the used shadow is expected to have an impact in the accuracy, but mainly in the resolution, of the reconstructed geometry. The radiometric form, even in the enhanced shadow carving approach, fails to capture complex geometries that are frequent in cultural artifacts, and thus it should be applied very cautiously.

47.3.10 Structure from Motion Structure from motion (SFM) is a photogrammetric, triangulation-based method that operates on a large set of photos of a static scene or of an object taken from various viewpoints, generally, with cameras of unknown characteristics and settings. SFM simultaneously estimates the location of object points (structure) and the pose of the camera (motion). As in all previously presented methods, a basic prerequisite for the application of SFM is that the object or scene be rigid. SFM has been intensely studied for more than 30 years within the computer vision community. The relevant bibliography is vast, and it has been stated that the papers devoted to SFM are more than those in any other area of computer vision [87].

1392

G. Pavlidis and A. Koutsoudis

Fig. 47.13 Example of application of Shape from shadow as a special case of structured light

At the heart of SFM is the solution of the point correspondence problem in a multiple view setting; that is, a reference point may be detected in multiple photos. The process includes the detection of image features, the matching of image features across photos, the creation of tracks from matches, and the solution of the SFM problem from the tracks. In the SFM context, “tracks” are defined as 3D coordinates of reconstructed points accompanied with a list of the corresponding 2D coordinates in the photos [88]. The SFM process has been significantly automated by using advanced feature extraction approaches such as SIFT [89] and SURF [90]. The original version of SFM produces sparse point clouds from point correspondences and is therefore of limited practical value as it leads to very low resolution 3D models. Today, SFM has been successfully combined with Multi-View Stereo (MVS) methods that are capable of producing dense point clouds. In essence, SFM is being used to estimate the camera parameters and pose that initializes MVS that ultimately produces the final dense point cloud of the measured objects. A comparison and evaluation of multi-view stereo reconstruction algorithms is provided in [91]. The hybrid SFM-MVS method has been very successful and

47

3D Digitization of Tangible Heritage

1393

Fig. 47.14 3D reconstruction using SFM-MVS

tends to become the most widespread and common photogrammetric approach adopted by experts in various domains. SFM-MVS can be of low cost with highquality (accuracy and resolution) digitization results when properly applied. The creation of 3D digital replicas of cultural heritage objects by SFM-MVS tends to become the most popular solution, surpassing laser scanning and structured light approaches. An example of a 3D model created by SFM-MVS is shown in Fig. 47.14. The figure shows the point cloud, the 3D mesh, and the final textured model. Practically, SFM-MVS is based on a set of complex and computationally demanding and memory hungry algorithms. The method is robust in small changes in light and intensity of the colors in the photos, and this makes it even more userfriendly for outdoor or indoor applications. As with any other method, SFM-MVS has its limits. Since it is based on point correspondences on object surfaces among photos, the quality of the 3D reconstruction is related to the existence of intense morphological features. Such morphological features are the frequent changes in intensity or chromaticity. The application of the method on reflective surfaces is greatly influenced by the use of particular filters, like polarizers, that minimize reflections and light scattering. Applying the method to featureless surfaces can lead to incorrect point matches among photos due to the uncertainty in the identification of the points correspondences, and ultimately, to the presence of intense noise in the data (thus in a low-quality 3D reconstruction), or to the creation of objects with severe deformations. Several studies have been carried out to address this problem, among which it has been proposed to project noise patterns on the surface of the measured objects, using a projection system, so as to enrich the surface with the required synthetic changes in intensity and chromaticity that would lead to accurate point correspondences [92]. In particular, research has already focused on the issues of featureless and specular surfaces, the image pre-processing for SFM-MVS optimization, and the selection of the proper acquisition setup [93–101]. Nowadays, the 3D data resulting from SFM-MVS systems are of high quality, and such systems are currently being used in demanding cultural digitization applications. As SFM-MVS solutions are of relatively low cost and can be automated and easily implemented by non-specialists, they are even more appealing. Of course, as with all methods, understanding how SFM-MVS works and what the limitations are, are very important factors for its proper implementation, both

1394

G. Pavlidis and A. Koutsoudis

during the data acquisition process and during the data processing for the creation of the 3D digital replica. On the other hand, the high quality of the data resulting from the correct application of this method renders SFM-MVS an efficient and successful solution that can serve a wide range of digitization projects.

47.4

The Influence of Object Characteristics

In every digitization project, of primary concern for the proper design of its implementation is the size and freedom of movement of the digitization subjects. This is also the main reason why as a first step in the identification of the project implementation requirements, objects are distinguished into movable and immovable. Movable can be any relatively small, easily transportable objects, whose digitization can either be done on-site or in special laboratories. Movable objects can be digitized in three dimensions using a variety of methods, such as laser beam triangulation, structured light scanning, or photogrammetry. Typical examples of movable cultural objects are statues, vases, utensils, jewelry, and folk art objects. On the other side, large statues, structures and buildings, architectural ensembles, urban areas of cultural and historical value, as well as archaeological sites and excavations are in the category of immovable objects, those that have to be digitized on-site (actually, they are the site). Typical examples of immovable monuments include the temples (e.g., Early Christian, Byzantine), archaeological excavations (terrestrial and underwater), as well as buildings of special architectural value (e.g., neoclassical) [102–105]. As mentioned earlier, applying a 3D digitization method includes the calculation of the positions and color of selected points on the surfaces of the objects by performing a sampling. In the 3D digital replica these points are represented by what are called the “vertices” that make up the outer surface of the object, which in several cases are defined in an arbitrary Cartesian coordinate system. Many digitization systems (such as laser scanners) produce 1:1 digital replicas, but this is not true for all systems. The aforementioned sampling is practically done without any control by the operator of the system, as the selection of the points on the surfaces of the objects cannot often be predetermined. Altogether, at the end of the measurement process, an arbitrarily large set of vertices is generated that is called the point cloud, which digitally represents the outer surface of the measured object and constitutes the raw digitization data. The goal during this phase is to create a dense point cloud (high resolution) of highly accurate measurements. As expected, this point cloud cannot be perfect, since in every measurement process there are errors related to the operating limits of the measurement system and external factors that affect each measurement. These errors are termed measurement noise. A point cloud is, nevertheless, an incomplete representation of a surface, as it represents only some of its points with discontinuities and gaps. Thus, point clouds need to pass through a series of processing steps for the final output of the 3D digital copy, or a 3D model. It is usually necessary to apply methods of filtering to eliminate potentially redundant points and to correct the positions of the points that have been misplaced, as well as to join the vertices into triangles or polygons to create a continuous surface.

47

3D Digitization of Tangible Heritage

1395

The creation of the final surface must (ideally) lead to a 3D digital object of closed manifold geometry (there is a more “relaxed” form of geometry that is very useful in 3D graphics, the one of “watertight” 3D models, which are particularly useful for interactive graphics applications and 3D printing. Manifolds are more strict structures, practically infeasible to attain in 3D digitization processes), on which the projection of the color information is integrated to complete the photorealistic representation. The variety of raw materials used in the creation of cultural objects, as well as their morphological complexity, are part of a wider set of features that are challenging for the various 3D digitization methods. Although, in principle, any digitization process should be performed with the highest possible resolution and accuracy using the most appropriate method, the quality of an object’s digitization is affected by the object itself. A typical material used in the literature to highlight the challenges that objects introduce to digitization methods is marble, since this material has two important properties: (a) translucence (b) heterogeneous surface roughness. These properties are inevitable sources of noise and affect the quality of the measurements [106]. The surface of a marble is structured by densely packed crystals. The variable density of the crystals causes a heterogeneity in the substrates and therefore alters the optical characteristics of the material. As light penetrates a marble surface, the unevenness of the substrates leads to light scattering through a series of refractions and reflections. This scattering is detected as bright concentric discs around the reflection point. Some types of marbles allow the diffusion of ambient light into their inner layers, thus revealing a series of particularly colorful features. As a result light detecting methods face serious difficulties in capturing the surface. Similar issues but of different nature arise in various materials, like ivory or fabric and fur. In addition, since most 3D digitization methods rely on optical measurements, highly reflective or black (and very dark) objects make it virtually impossible to succeed in capturing their surface. Objects like mirrors, glasses, jewelry, or blackfigure or red-figure pottery are typical examples in this case. Things get even more complicated for morphologically complex objects with lots of cavities and holes, which a light beam cannot reach or its reflection cannot be detected. The morphological complexity that is typically inherent in cultural objects imposes other constraints also, like the need for safe distances and handling during the digitization. This might be a significant challenge in cases in which the safety precautions prevent the digitization system from effectively measuring reflected light from the surface of the measured object. To summarize, a large number of objects display highly complex morphology and surface properties that either cause problems or render 3D digitization infeasible with technologies currently available [107, 108].

47.4.1 Object Size and 3D Digitization As one of the most important characteristics of the cultural objects that influence the proper selection of a method and the process of 3D digitization, the following

1396

G. Pavlidis and A. Koutsoudis

paragraphs focus on the categories of object scales and the corresponding applicable digitization methods.

Microscopic Objects Today, it is possible to digitize really small objects using various methods. Techniques such as Atomic Force Microscopy and Stereoscopic Scanning Electron Microscopy enable the detection and recording of characteristics on the scale of tens of nanometers (nm). Other techniques, such as Confocal Microscopy and White Light Scanning Interferometry, provide a detection resolution on the scale of hundreds of nanometers. Although such methods match the requirements for digitizing microscopic objects in high resolution and accuracy, they rely on particularly costly specialized equipment and can be applied only to very small surfaces (typically of areas smaller than 1 mm2). Apparently, due to these limitations, the practical value of those methods is questioned, and their application should be selected after careful balancing of the required digitization outcome and the available resources. Applications of Stereoscopic Scanning Electron Microscopy are rather limited. In this method, the objects have to be put into vacuum, and a special metal coating is usually required. Apparently, this is a “destructive” technique and is therefore inappropriate for most application in cultural heritage. There are, of course, cases such as broken ceramics, fossils, or glass objects where it could be applied. Confocal Microscopy is typically applied in the domain of life sciences and gives the best results when used in conjunction with techniques such as fluorescence. It yields satisfactory results for translucent objects, allowing the recording of internal information. It could be selected as a method for transparent or translucent objects. White Light Scanning Interferometry can provide good digitization at a significantly lower cost. Commercial systems allow for the measurement of 3D surfaces in the sub-micrometer (μm) scale and could be considered particularly useful for the thorough study of surfaces. The method does not require contact or any special surface preparation. Atomic Force Microscopy is the youngest in the category. Its principle lies in the use of a very small measuring system, with a size of no more than ten atoms, which is placed very close to the measured surface without being in contact. Intra-atomic forces draw the metering system to the surface. A microscope detects the measurement system in its motion along the measured surface, and the measurement of the geometry is based on calculations of the exerted pulling force (between the measuring system and the surface). The method can provide very high accuracy, even under one nanometer in special cases. The reasonable question is if there is something on the scale of the microscopic objects that would be interesting to detect and record. In this scale, the micro-texture of a surface or its roughness can be recovered, which may have a significant impact on the overall shape of the object. In addition, changes in surface roughness could be characteristic of the surface condition (e.g., corrosion). Indeed, on the micro scale there is important information. Thus, of the four techniques mentioned, Atomic Force Microscopy and White Light Scanning Interferometry can contribute to the 3D digitization of tangible heritage. Nevertheless, there are specific drawbacks to be

47

3D Digitization of Tangible Heritage

1397

kept in mind, like the limitation to very small surfaces, the significant complexity and costs, the expertise required, and the laboratory conditions required.

Small Objects Small are considered objects of up to 100  100 mm (10  10 cm). Undoubtedly in this category popular techniques are those based on Laser scanning. There is currently a large variety of commercially available laser scanning devices for this category of objects. The basic principle of operation is relatively simple and is based on the concept of triangulation: a laser light source shines on the measured surface, and its reflection is detected by an optical sensor. Given that the area to be measured is limited to 100  100 mm, these systems achieve a measurement accuracy of the order of some micrometers. In the digitization of small objects, many other methods can be applied. Methods like Shape from structured light, Shape from silhouette, and Structure from motion can provide high-quality digitization results even under no special conditions or limitations. Medium-Sized Objects The category of medium-sized objects is a category of rather interesting objects, where most of the available 3D digitization methods can be effectively applied. Objects are more than 10 cm in size and can reach up to 1 or 2 m. In terms of cultural heritage, objects like statues of natural size can be included in this category. Apart from Laser scanning, structured light techniques are rather popular in this category of objects. The operation of these techniques is based on the principle of triangulation. As will be explained in following paragraphs, structured light methods are lower cost triangulation approaches that can achieve remarkable digitization results. Typically, a projection system projects a light pattern onto the measured surface, and an imaging system records the lighted surface. The deformations of the pattern correspond to the surface geometry, which is extracted by triangulation. Commercial systems in this category are capable of achieving an accuracy of about 1/100 of the interval between the fringes of the pattern. If, for example, this interval is 1 cm, then the expected measurement accuracy is of the order of 0.1 mm. One important feature of structured light techniques is that their accuracy is scalable with the distance. Since a large variety of object sizes is included in this category of objects, depending on the digitization project requirements and limitations, a given digitization system needs to change its distance from the objects, thus increasing or decreasing the effectiveness of its imaging system to accurately recognize the projected pattern and precisely estimate the distance. Important drawbacks of structured light systems include (a) limitations in detecting very dark surfaces, (b) issues in detecting glossy objects with reflective surfaces, (c) negative influences of surface patterns or paintings, and (d) limitations in capturing surfaces with intense morphological complexity. A rather interesting family of techniques that have a great potential in this category of objects is that of holography. Traditional holography is here since the mid-twentieth century but has not yet been able to reach the market with commercial products, due to inherent characteristics, such as the requirement for liquid chemicals

1398

G. Pavlidis and A. Koutsoudis

for the exposure of images. However, the trend might change with computer assisted holography. Holography draws its principle from the nature of light as an electromagnetic way, which has two important properties, the intensity and the frequency and an important implication of time, the phase (a measure of synchronicity). A black and white (more correctly, a grayscale) image created by captured light is a representation of light intensity. A color image is an imprint of the intensity and the frequency of captured light, as color in light is the observable phenomenon of its frequency. A hologram is an imprint of captured light waves, which convey intensity, frequency, and phase information. As light waves coming from different points of the observed object have traveled different distances, they reach the detector with shifts in their phase. Apparently, this phase shift seems to encode the difference in the distance that light travels; thus, the object’s geometry may be recovered by phase information. Today, there are two approaches to computer assisted holography, Digital holography and Computer generated holography. Both have significant prospects in capturing (and displaying) 3D objects. Main advantages of holographic digitization methods include the high resolution and accuracy and the significant longevity of the data, as they are not tied to any contemporary technology but are purely based on light, which is a universal standard. Another family of techniques that can be studied for the particular scale of medium-sized objects is tomography. Tomography is in essence the reconstruction of a volume by dense 2D sectional images, and is typically found in applications in medicine and aerospace. Tomography is capable of high accuracy measurements, especially for small objects. Its main advantage and unique characteristic among all other digitization approaches lies in the ability to capture and reconstruct an internal structure. A widespread form of tomography is using X-rays, but the family of methods also includes approaches of visual and acoustic tomography. When tomography is applied to living organisms, it is based on detecting the different levels or rate of absorption of a particular radiation by the various materials of the bodies, such as bones, muscles, liquids, gases, etc. In general, however, the lack of differentiation in the internal structure of cultural objects renders tomographic methods less interesting. Nevertheless, there are cases in which tomography is the most appropriate digitization method, like in the case of sarcophagi. Recently, following significant advances in computer technology, computer vision research, and efficient implementation of powerful algorithms, another 3D digitization method that has been successfully applied in medium-sized objects is Structure from motion. In fact, as Structure from motion has been integrated with Multiple-view stereo, its modern approach achieves high resolution and accuracy 3D digitization results. The method is based on the solution of the inverse problem in photography, which consists in recovering 3D structure from 2D photos. Multiple photos are being used that cover the entire surface of the measured object, and special computer vision algorithms take over to solve the inverse problem step by step. The method is implemented exclusively in software, which is fed by the usually large set of photos and can result in very high-quality 3D digital replicas. Studies have shown that if the method is used with knowledge of its requirements and

47

3D Digitization of Tangible Heritage

1399

limitations, it can achieve results comparable to those attained by much more expensive laser triangulation-based methods (see, e.g., [109]).

Large Objects and Monuments This category includes objects larger than 2 m, which may even reach 100 m or more. While there are many techniques for measuring medium objects, there is a lack of a variety of methods in this class. There are basically two basic families of methods effectively applicable in this class. In the first family, the methods are based on the idea of optical radar. In the simplest form a source emits laser rays, which are either amplitude or frequencymodulated. The emitted beam is reflected from the target and detected by the measurement system. By examining the modulation state of the detected beam, it is possible to calculate the time taken by the beam to travel the corresponding distance from the transmitter to the object and back to the receiver. Knowing the speed of light in the air, it is possible to calculate the distance traveled. These systems are known as Time of Flight scanners (TOF). In this category there are commercially available devices capable of measuring multi-million points and composing an integrated 3D scene. They provide an accuracy in the scale of some millimeters for various operating distances, but with an accuracy and resolution that is a function of the distance. The obvious drawback of these systems is that they make pointbased measurements and since large scenes are being scanned, a considerable amount of time is needed on site to collect all the necessary data. The second family of techniques is photogrammetry. The most basic of these techniques is the well-known Guided stereo-photogrammetry. In this method reference points are used, which are defined on the surface of the measured objects, and at least one pair of photographs from different angles is obtained. By matching the reference points, it is possible to retrieve 3D geometry information by measuring the distance from the capturing system. The method obviously gives better results when there are flat surfaces with few points of reference, or generally, in the scene, [110]. Recently, with the significant advancement of Structure from Motion (which can be considered a member of this family) and the advancement of the technology of remotely operated vehicles like the UAVs, it has made possible and rather efficient to couple UAVs with Structure from motion and tackle large object digitization. There are paradigms of independent and large EU projects (like project 3D-Icons http://3dicons-project.eu) which have created remarkable 3D reconstructions by coupling terrestrial and aerial Structure from motion.

47.5

Conclusion

This chapter focuses on the presentation of approaches and methods for the 3D digitization of tangible heritage. To this end, it first introduces the main concepts of digitization, borrowing the foundations from the electrical engineering domain. In this view, digitization is the recording of quantized values of discrete measurements of quantities of the physical world, a discrete sampling of reality. In tangible cultural

1400

G. Pavlidis and A. Koutsoudis

heritage, there is multifaceted motivation for the application of 3D digitization, which, basically, consists in recording the geometric structure and spectral response of artifacts and monuments. Facilitating the study, research, comparative assessment, preservation, safeguarding, dissemination, education, and duplication are among the most pronounced motivations for the application of 3D digitization in tangible heritage. The chapter introduces the available 3D digitization methods along with the basic ideas on their principles, their strengths, and weaknesses. Last but not least, it also discusses the influence of object characteristics in the selection and application of the most appropriate methods and particularly on how the objects’ size is a decisive factor in cultural heritage 3D digitization projects.

References 1. Cutting JE, Vishton PM (1995) Perceiving layout and knowing distances: the integration, relative potency, and contextual use of different information about depth. In: Perception of space and motion. Elsevier, pp 69–117 2. Beraldin J-A, Blais F, Cournoyer L, Godin G, Rioux M, Taylor J (2004) Active 3D sensing for heritage applications. Bar International Series 1227. TEMPUS REPARATSM, pp 340–343 3. Boehler W, Marbs A (2002) 3D scanning instruments. Proceedings of the CIPA WG 6, pp 9–18 4. Forest J, Salvi J, Cabruja E, Pous C (2004) Laser stripe peak detector for 3D scanners. A FIR filter approach. In: Pattern recognition, 2004. ICPR 2004. Proceedings of the 17th international conference on, vol 3, pp 646–649 5. Cortelazzo GM, Marton F (1999) About modeling cultural heritage objects with limited computers resources. In: Image analysis and processing, 1999. Proceedings. international conference on, pp 848–853 6. Rioux M (1994) Digital 3D imaging: theory and applications. Videometrics III 2350:2–16 7. Curless B, Levoy M (1995) Better optical triangulation through spacetime analysis. In: ICCV, p 987 8. Sobieszek RA (1980) Sculpture as the sum of its profiles: Francois Willeme and photosculpture in France, 1859–1868. Art Bull 62(4):617–630 9. Baumgart BG (1974) Geometric modeling for computer vision 10. Martin WN, Aggarwal JK (1983) Volumetric descriptions of objects from multiple views. IEEE Trans Pattern Anal Mach Intell 2:150–158 11. Zhang Z (1998) Modeling geometric structure and illumination variation of a scene from real images. In: Computer vision, 1998. Sixth international conference on, pp 1041–1046 12. Tosovic S, Sablatnig R, Kampel M (2002) On combining shape from silhouette and shape from structured light 13. Laurentini A (1994) The visual hull concept for silhouette-based image understanding. IEEE Trans Pattern Anal Mach Intell 16(2):150–162 14. Sablatnig R, Tosovic S, Kampel M (2003) Next view planning for shape from silhouette, computer vision (CVWW’03), Czech Pattern Recognition Society, pp 77–82 15. Baumberg A, Lyons A, Taylor R (2005) 3D SOM – a commercial software solution to 3D scanning. Graph Model 67(6):476–495 16. Potmesil M (1987) Generating octree models of 3D objects from their silhouettes in a sequence of images. Comp Vision Graph Image Process 40(1):1–29 17. Noborio H, Fukuda S, Arimoto S (1988) Construction of the octree approximating a threedimensional object by using multiple views. IEEE Trans Pattern Anal Mach Intell 10(6):769–782 18. Ahuja N, Veenstra J (1989) Generating octrees from object silhouettes in orthographic views. IEEE Trans Pattern Anal Mach Intell 11(2):137–149

47

3D Digitization of Tangible Heritage

1401

19. Matsumoto Y, Terasaki H, Sugimoto K, Arakawa T (1997) A portable three-dimensional digitizer. In: 3-D digital imaging and modeling, 1997. Proceedings. International conference on recent advances in, pp 197–204 20. Lensch HP, Heidrich W, Seidel H-P (2001) A silhouette-based algorithm for texture registration and stitching. Graph Model 63(4):245–262 21. Loop C, Zhang Z (1999) Computing rectifying homographies for stereo vision. In: Proceedings. 1999 IEEE Computer Society conference on computer vision and pattern recognition (Cat. No PR00149), vol 1, pp 125–131 22. Scharstein D, Szeliski R (2002) A taxonomy and evaluation of dense two-frame stereo correspondence algorithms. Int J Comput Vis 47(1–3):7–42 23. Bertozzi M, Broggi A, Conte G, Fascioli A (2002) Stereo-vision system performance analysis. In: Enabling technologies for the PRASSI autonomous robot, pp 68–73 24. Szeliski R (2011) Computer vision – algorithms and applications. Springer, London 25. Zanuttigh P, Marin G, Dal Mutto C, Dominio F, Minto L, Cortelazzo GM (2016) Operating principles of structured light depth cameras. In: Time-of-flight and structured light depth cameras. Springer, pp 43–79 26. Salvi J, Pages J, Batlle J (2004) Pattern codification strategies in structured light systems. Pattern Recogn 37(4):827–849 27. Rocchini C, Cignoni P, Montani C, Pingi P, Scopigno R (2001) A low cost 3D scanner based on structured light. Comput Graphics Forum 20:299–308 28. Horn E, Kiryati N (1999) Toward optimal structured light patterns. Image Vis Comput 17(2): 87–97 29. Chen C-S, Hung Y-P, Chiang C-C, Wu J-L (1997) Range data acquisition using color structured lighting and stereo vision. Image Vis Comput 15(6):445–456 30. Wust C, Capson DW (1991) Surface profile measurement using color fringe projection. Mach Vis Appl 4(3):193–203 31. Horn BK (1975) Obtaining shape from shading information. In: The psychology of computer vision. McGraw-Hill Book Co., New York, pp 115–155 32. Brooks MJ (1983) Two results concerning ambiguity in shape from shading. AAAI 83 33. Oliensis J (1991) Shape from shading as a partially well-constrained problem. CVGIP: Image Underst 54(2):163–183 34. Oliensis J (1991) Uniqueness in shape from shading. Int J Comput Vis 6(2):75–104 35. Rouy E, Tourin A (1992) A viscosity solutions approach to shape-from-shading. SIAM J Numer Anal 29(3):867–884 36. Belhumeur PN, Kriegman DJ, Yuille AL (1999) The bas-relief ambiguity. Int J Comput Vis 35(1):33–44 37. Durou J-D, Piau D (2000) Ambiguous shape from shading with critical points. J Math Imaging Vision 12(2):99–108 38. Prados E, Faugeras OD (2003) “Perspective shape from shading” and viscosity solutions.. In: ICCV, vol 3, p 826 39. Prados E, Faugeras O (2006) Shape from shading. In: Handbook of mathematical models in computer vision. Springer, pp 375–388 40. Zhang R, Tsai P-S, Cryer JE, Shah M (1999) Shape-from-shading: a survey. IEEE Trans Pattern Anal Mach Intell 21(8):690–706 41. Horn BK, Sjoberg RW (1978) Calculating the reflectance map 42. Wu Y (2003) Radiometry, BRDF and photometric stereo. Evanston 43. Woodham RJ (1980) Photometric method for determining surface orientation from multiple images. Opt Eng 19(1):139–144 44. U. S. N. B. of Standards, Nicodemus FE (1977) Geometrical considerations and nomenclature for reflectance, vol 160. US Department of Commerce, National Bureau of Standards 45. Christensen PH, Shapiro LG (1994) Three-dimensional shape from color photometric stereo. Int J Comput Vis 13(2):213–227 46. Hertzmann A, Seitz SM (2005) Example-based photometric stereo: shape reconstruction with general, varying BRDFS. IEEE Trans Pattern Anal Mach Intell 27(8):1254–1264

1402

G. Pavlidis and A. Koutsoudis

47. Verma CS, Wu M-J (2009) Photometric stereo 48. Fulcoly DO, Kalamaroff KI, Chun F (2012) Determining basic satellite shape from photometric light curves. J Spacecr Rocket 49(1):76–82 49. Yeh C-K, Li F, Pastorelli G, Walton M, Katsaggelos AK, Cossairt O (2017) Shape-fromshifting: uncalibrated photometric stereo with a mobile device. In: 2017 IEEE 13th international conference on e-Science (e-Science), pp 551–558 50. Aloimonos J, Chou PB (1985) Detection of surface orientation and motion from texture: 1. The case of planes. Department of Computer Science, University of Rochester 51. Aloimonos J (1988) Shape from texture. Biol Cybern 58(5):345–360 52. Kanatani K-i, Chou T-C (1989) Shape from texture: general principle. Artif Intell 38(1):1–48 53. Forsyth DA (2002) Shape from texture without boundaries. In: European conference on computer vision, pp 225–239 54. Kanatani K-I (2012) Group-theoretical methods in image understanding, vol 20. Springer Science & Business Media 55. Gibson JJ (1950) The perception of the visual world. Houghton Mifflin 56. Pentland A (1985) A new sense for depth of field. In: Proceedings of the ninth international joint conference on artificial intelligence, Los Angeles, pp 988–994 57. Pentland AP (1987) A new sense for depth of field. IEEE Trans Pattern Anal Mach Intell 4: 523–531 58. Grossmann P (1987) Depth from focus. Pattern Recogn Lett 5(1):63–69 59. Subbarao M (1987) Direct recovery of depth-map II: a new robust approach 60. Darrell T, Wohn K (1988) Pyramid based depth from focus. In: Proceedings CVPR’88: the Computer Society conference on computer vision and pattern recognition, pp 504–509 61. Ohta T, Sugihara K, Sugie N (1988) A method for image composition using image variance. Trans IECE 66(10):1245–1246 62. Kaneda K, Wakasu Y, Nakamae E, Tazawa E (1988) A method of pan-focused and stereoscopic display using a series of optical microscopic images. In: Proceedings of fourth symposium on image sensing technologies in industry, pp 189–194 63. Nayar SK (1989) Shape from focus 64. Nakagawa Y, Nayer SK (1992) Method of detecting solid shape of object with autofocusing and image detection at each focus level, US Patent 5,151,609, Google Patents 65. Ens J, Lawrence P (1993) An investigation of methods for determining depth from focus. IEEE Trans Pattern Anal Mach Intell 15(2):97–108 66. Xiong Y, Shafer SA (1993) Depth from focusing and defocusing. In: Proceedings of IEEE conference on computer vision and pattern recognition, pp 68–73 67. Favaro P (2000) Shape from focus/defocus. Washington University 68. Das S, Ahuja N (1995) Performance analysis of stereo, vergence, and focus as depth cues for active vision. IEEE Trans Pattern Anal Mach Intell 17(12):1213–1219 69. Chaudhuri S, Rajagopalan A (1999) Depth from defocus: a real aperture imaging approach. Springer Science & Business Media 70. Bailey SW, Echevarria JI, Bodenheimer B, Gutierrez D (2015) Fast depth from defocus from focal stacks. Vis Comput 31(12):1697–1708 71. Suwajanakorn S, Hernandez C, Seitz SM (2015) Depth from focus with your mobile phone. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 3497–3506 72. Tang H, Cohen S, Price B, Schiller S, Kutulakos KN (2017) Depth from defocus in the wild. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2740–2748 73. Hazirbas C, Soyer S, Staab M, Leal-Taixe L, Cremers D, Mayer N, Ilg E, Fischer P, Dosovitskiy A, Brox T (2018) Deep depth from focus. IJCV 74. Shafer SA, Kanade T (1983) Using shadows in finding surface orientations. Comput Vis Graph Image Process 22(1):145–176 75. Hatzitheodorou M, Kender JR (1988) An optimal algorithm for the derivation of shape from shadows. In: Proceedings CVPR’88: the Computer Society conference on computer vision and pattern recognition, pp 486–491

47

3D Digitization of Tangible Heritage

1403

76. Raviv D, Pao Y-H, Loparo KA (1989) Reconstruction of three-dimensional surfaces from two-dimensional binary images. IEEE Trans Robot Autom 5(5):701–710 77. Langer MS, Dudek G, Zucker SW (1995) Space occupancy using multiple shadow images. In: Proceedings 1995 IEEE/RSJ international conference on intelligent robots and systems. Human robot interaction and cooperative robots, vol 1, pp 285–290 78. Yang DK-M (1996) Shape from darkness under error. PhD thesis. Columbia University 79. Daum M, Dudek G (1998) On 3-D surface reconstruction using shape from shadows. In: Proceedings. 1998 IEEE Computer Society conference on computer vision and pattern recognition (Cat. No. 98CB36231), pp 461–468 80. Savarese S, Rushmeier H, Bernardini F, Perona P (2001) Shadow carving. In: Proceedings eighth IEEE international conference on computer vision. ICCV 2001, vol 1, pp 190–197 81. Savarese S, Rushmeier H, Bernardini F, Perona P (2002) Implementation of a shadow carving system for shape capture. In: Proceedings. First international symposium on 3D data processing visualization and transmission, pp 12–23 82. Savarese S, Andreetto M, Rushmeier H, Bernardini F, Perona P (2007) 3d reconstruction by shadow carving: theory and practical evaluation. Int J Comput Vis 71(3):305–336 83. Yu Y, Chang JT (2005) Shadow graphs and 3D texture reconstruction. Int J Comput Vis 62(1–2):35–60 84. Abrams A, Miskell K, Pless R (2013) The episolar constraint: monocular shape from shadow correspondence. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 1407–1414 85. Bouguet J-Y, Perona P (1998) 3D photography on your desk. In: Computer vision, 1998. Sixth international conference on, pp 43–50 86. Kawasaki H, Furukawa R (2009) Shape reconstruction and camera self-calibration using cast shadows and scene geometries. Int J Comput Vis 83(2):135 87. Oliensis J (2000) A critique of structure-from-motion algorithms. Comput Vis Image Underst 80(2):172–214 88. Furukawa Y, Hernandez C (2015) Multi-view stereo: a tutorial. Found Trends Comput Graph Vis 9(1–2):1–148 89. Lowe DG (1999) Object recognition from local scale-invariant features. In: Computer vision, 1999. The proceedings of the seventh IEEE international conference on, vol 2, pp 1150–1157 90. Bay H, Tuytelaars T, Van Gool L (2006) Surf: speeded up robust features. In: European conference on computer vision, pp 404–417 91. Seitz SM, Curless B, Diebel J, Scharstein D, Szeliski R (2006) A comparison and evaluation of multi-view stereo reconstruction algorithms. In: 2006 IEEE Computer Society conference on computer vision and pattern recognition (CVPR’06), vol 1, pp 519–528 92. Koutsoudis A, Ioannakis G, Vidmar B, Arnaoutoglou F, Chamzas C (2015) Using noise function-based patterns to enhance photogrammetric 3D reconstruction performance of featureless surfaces. J Cult Herit 16(5):664–670 93. Baltsavias EP (1991) Multiphoto geometrically constrained matching. PhD thesis, Columbia University 94. Baltsavias E, Li H, Mason S, Stefanidis A, Sinning M (1996) Comparison of two digital photogrammetric systems with emphasis on DTM generation: case study glacier measurement. Int Arch Photogramm Remote Sens 31:104–109 95. Mallick SP, Zickler TE, Kriegman DJ, Belhumeur PN (2005) Beyond lambert: reconstructing specular surfaces using color. In: 2005 IEEE Computer Society conference on computer vision and pattern recognition (CVPR’05), vol 2, pp 619–626 96. Remondino F, El-Hakim S, Gruen A, Zhang L (2008) Development and performance analysis of image matching for detailed surface reconstruction of heritage objects. IEEE Signal Process Mag 25(4):55–65 97. Guidi G, Gonizzi S, Micoli L (2014) Image pre-processing for optimizing automated photogrammetry performances. In: ISPRS Technical Commission V symposium, vol 2, pp 145–152

1404

G. Pavlidis and A. Koutsoudis

98. MacDonald L, Hindmarch J, Robson S, Terras M (2014) Modelling the appearance of heritage metallic surfaces. Int Arch Photogramm Remote Sens Spat Inf Sci 45:371–377 99. Gaiani M, Remondino F, Apollonio F, Ballabeni A (2016) An advanced preprocessing pipeline to improve automated photogrammetric reconstructions of architectural scenes. Remote Sens 8(3):178 100. Nikolov I, Madsen C (2016) Benchmarking close-range structure from motion 3D reconstruction software under varying capturing conditions. In: Euro-Mediterranean conference, pp 15–26 101. Ioannakis G, Koutsoudis A, Arnaoutoglou F, Kiourt C, Chamzas C (2017) On structure-frommotion application challenges: good practices. Int J Comput Methods Herit Sci (IJCMHS) 1(2):47–57 102. Patias P (1999) Photogrammetric survey and documentation of architectural monuments and archaeological sites 103. Tziavos I, Spatalas S (2004) Urban design applications and topographic surveys 104. Daniil M (2009) Topography – topographic mapping of space 105. Vlachos D (1998) Principles and methods of topographic recording 106. Godin G, Rioux M, Beraldin J-A, Levoy M, Cournoyer L, Blais F (2001) An assessment of laser range measurement on marble surfaces. In: 5th conference on optical 3D measurement techniques, vol 3 107. Hawkins T, Cohen J, Debevec P (2001) A photometric approach to digitizing cultural artifacts. In: Proceedings of the 2001 conference on virtual reality, archeology, and cultural heritage, pp 333–342 108. Marbs A (2002) Experiences with laser scanning at i3mainz. In: Proceedings of the CIPA WG6 international workshop on scanning for cultural heritage recording. http://www.isprs.org/ commission5/workshop 109. Koutsoudis A, Vidmar B, Ioannakis G, Arnaoutoglou F, Pavlidis G, Chamzas C (2014) Multiimage 3D reconstruction data evaluation. J Cult Herit 15(1):73–79 110. Streilein A, van den Heuvel FA (1999) Potential and limitation for the 3D documentation of cultural heritage from a single image. In: Proceedings XVII CIPA symposium October, vol 99

eXtended Reality for Cultural Heritage

48

Ronan Gaugne, Jean-Baptiste Barreau, Flavien Le´cuyer, The´ophane Nicolas, Jean-Marie Normand, and Vale´rie Gouranton

Contents 48.1 48.2 48.3

48.4

48.5

48.6

Introduction to eXtended Reality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . eXtended Reality for Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.3.1 “See” Axis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.3.2 “Manipulate” Axis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.3.3 “Share” Axis Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3D Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.4.1 Digitization and Data Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.4.2 Graphical Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XR Design and Implementation for Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.5.1 XR Systems for Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.5.2 XR Interaction and Collaboration in Cultural Heritage . . . . . . . . . . . . . . . . . Evaluation and Use by Cultural Heritage Experts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.6.1 Study of a Cremation Urn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.6.2 Reconstitution of an Eighteenth-Century Ship . . . . . . . . . . . . . . . . . . . . . . . . . . 48.6.3 Process of Analysis of a Gallic Grave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1406 1407 1409 1410 1410 1411 1411 1412 1413 1416 1416 1418 1424 1424 1426 1428

R. Gaugne (*) Univ Rennes, Inria, CNRS, IRISA, Rennes, France e-mail: [email protected] J.-B. Barreau CNRS, Université Paris 1 Panthéon-Sorbonne, Paris, France e-mail: [email protected] F. Lécuyer · V. Gouranton (*) Univ Rennes, INSA Rennes, Inria, CNRS, IRISA, Rennes, France e-mail: fl[email protected]; [email protected] T. Nicolas (*) Inrap, UMR 8215 Trajectoires, Rennes, France e-mail: [email protected] J.-M. Normand Ecole Centrale de Nantes, AAU UMR 1563, Nantes, France e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_48

1405

1406

R. Gaugne et al.

48.6.4 Annotation of Megalithic Art with Augmented Reality . . . . . . . . . . . . . . . . 48.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1430 1431 1432

Abstract

3D data production techniques, although increasingly used by archaeologists and Cultural Heritage practitioners, are most often limited to the production of 2D or 3D images. Beyond the modes of visualization of these data, it is necessary to wonder about their possible interactions and uses. Virtual Reality, Augmented Reality, and Mixed Reality, collectively known as eXtended Reality, XR, or Cross Reality, make it possible to envisage natural and/or complex interactions with 3D digital environments. These are physical, tangible, or haptic (i.e., with effort feedback) interactions, which can be understood through different modalities or metaphors, associated with procedures or gestures. These interactions can be integrated by archaeologists in the “operating chain” of an operation (from the ground to the study phase), or be part of a functional reconstitution of the procedures and gestures of the past, in order to help understand an object, a site, or even a human activity. The different case studies presented in this chapter result from collaborations between archaeologists, historians, and computer scientists. They illustrate different interactions in 3D environments, whether they are operational (support for excavation processes) or functional (archaeological objects, human activities of the past).

48.1

Introduction to eXtended Reality

eXtended Reality (XR) is generally presented as an encompassing concept that integrates technological platforms, 3D digital content, as well as the user experience to view and interact with digital and/or physical objects in a real and/or digital environment. XR can also be seen as a generic term that includes Virtual Reality (VR), Augmented Reality (AR), and Mixed Reality (MR), the latter being generally seen as a generalization of both VR and AR (see Fig. 48.1). Virtual reality is often considered solely as a field of computer science in relation to interactive digital 3D worlds. It actually holds a special position in the usual scientific scheme by coupling humanities sciences with engineering. According to Fuchs et al. [2] “The purpose of virtual reality is to allow a sensorimotor and cognitive activity for a person (or persons) in a digitally created artificial world, which can be imaginary, symbolic or a simulation of certain aspects of the real

Fig. 48.1 The Milgram continuum [1]

48

eXtended Reality for Cultural Heritage

1407

world.” This assertion positions the user and his/her activity at the center of VR. The user is immersed through his/her mind, senses, and body. Immersion is the “state (perceptual, mental, emotional) of a subject when one or more senses are isolated from the outside world and are fed solely by information from the computer” [3]. For mixed reality applications, immersion can be an important factor, as it allows the user to focus more on the artificial content displayed. Related to VR, Augmented Reality is a domain of applications that combines the real and the virtual, in real time, and gives the appearance that virtual and real objects “coexist in the same three-dimensional world” [4]. AR increases the visualization of reality with digital data and allows the user to interact with both environments (real and/or digital). For both virtual and augmented reality, the environment needs to be reactive, in order to change according to the user’s actions. To make it possible, Mixed Reality applications need to be real-time, i.e., that actions of the user need to be taken into account and transformed into actions on the digital content without the user being able to notice any delay. These different areas of Mixed Reality are presented in the Milgram continuum [1] (Fig. 48.1) in which one starts from real environments from one end of the continuum and go to pure Virtual Environments (VE) at the other end, while passing through Augmented Reality. As this is a continuum, many modalities co-exist in it. For instance, VR can be more or less immersive, depending on the kind of device used: a Head-Mounted Display (HMD) will enclose the user more strongly in the VE, making it farther on the VR side of the continuum than, for instance, a single screen. The different modalities existing in the spectrum of MR each have their own advantages and drawbacks, making each of them more or less adapted depending on the application.

48.2

eXtended Reality for Cultural Heritage

In archaeology, scientific processes based on VR, MR, or AR have known a tremendous increase in recent years. Virtual archaeology was first introduced by Reilly in 1990 [5] and was initially presented for excavation recording and virtual reexcavation using multimedia technologies. In a similar way, Krasniewicz, in 2000, proposed a 360 visualization infrastructure to help archaeologists in their research work [6]. In this case, virtual archaeology was not used to restore knowledge, but to acquire new knowledge. In her works, Pujol-Tost also discussed the importance of the chosen representation for the 3D models in everywhere [7]. According to new trends in the domain, Forte [8], one of the mainstays of virtual archaeology, suggests to replace the terminology related to a “reconstitution of the past,” by the expression “Cyber Archaeology” relying on a “simulation of the past.” The distinction between “cyber archaeology” and “virtual archaeology” is justified by the extensive use of the term “virtual archaeology” for applications dedicated to visualization purposes. Hence, “cyber archaeology” makes the archaeologist more active in the process.

1408

R. Gaugne et al.

Christou et al. [9] used an immersive CAVE-like structure combined with haptic devices and 3D sound for pedagogic purposes in a museum exhibition, but also as a tool for research. The Archave project [10] integrated also a CAVE-like structure and proposed tools for archaeologists to study historical sites. Another interesting work [11] presents a tool where archaeologists collaborate remotely on shared virtual objects through realistic avatars reconstructed from 3D cameras. Their virtual reality framework proposed a rich toolkit of interaction features, including navigation, measurement, lighting, and dragging. However, due to real-time 3D capture and rendering of users, the visualization was restricted to small image resolution (320  240 pixels) to ensure fluid rendering (25 FPS). XR is widely used for Cultural Heritage valorization but remains seldom used for scientific purposes. We believe that fields of XR and Cultural Heritage may enjoy mutual benefits triggered by questions such as: 1. How can XR, considered in its whole scope from science and technology to human sciences and natural sciences, benefit the study of Cultural Heritage? 2. What are the specific challenges brought by Cultural Heritage? Elements of reciprocal challenges are presented in Fig. 48.2. They are centered on three key notions: perception, interaction, and collaboration. Indeed, it is crucial to determine the relevant perception Cultural Heritage experts have of their environment and material of study, and to propose adapted modalities of perception in XR, taking into account the specificity of this particular family of users. To this aim, we have to identify the operational processes existing in the Cultural Heritage domain, in term of analysis, treatment, or conservation, to improve these processes and to imagine new processes based on XR. In the same way, it is also fundamental to understand interaction needs required by Cultural Heritage experts, to identify the contexts of interaction to consider, and to determine the purposes of interaction. Obviously, 1:1 interaction is important, but it may also be interesting to consider different scales of interaction, and even multi-scale. For example, in the case of an excavation site, it is required to manage the scale of the whole site that can be very

Fig. 48.2 Reciprocal challenges between XR and CH

48

eXtended Reality for Cultural Heritage

1409

large and the scale of artifacts discovered on the site. Depending of the context, both tangible and intangible interactions must be considered, for example, to enable a physical interaction with an artifact [12], or to represent human activities in a context of intangible heritage [13]. Finally, Cultural Heritage experts need to collaborate and share their knowledge, which is related to collaborative virtual environments [8] and environment enrichment with annotations [14]. Immersion, particularly in 1:1 virtual environments (i.e., virtual environments represented at their real scale), allows Cultural Heritage experts such as archaeologists and historians to evaluate symbolic or cultural roles of architectural buildings. Moreover, Cultural Heritage experts can be placed in situations to validate specific activities, ranging from displacements within environments to more complex interactions by using haptic devices to evaluate the physical feasibility or the coherency of a task. This aspect strongly pertains to ergonomics and musculoskeletal activities, which are widely developed in the context of XR, especially in VR. As part of a scientific process, historical reconstitution in XR should aim at providing a meaningful environment for Cultural Heritage experts, which corresponds to the following scientific problems: • Is the representation of the historical universe credible for Cultural Heritage experts? Realism and objective credibility are diametrically opposed issues. Objective credibility is an intrinsic quality of the reconstitution depending on the expert’s perception. Just as an X-ray picture has functional credibility for a medical practitioner, virtual Cultural Heritage experts’ models and environments must be designed to ensure this credibility. As Hermon and Nikodem developed [15], 3D modeling can only be used for archaeology under the conditions of data transparency and evaluability of the reconstruction’s accuracy. • Are Cultural Heritage experts able to evaluate hypotheses in the virtual reconstitution? Beyond the credibility of the reconstitution, the ability of an expert to acquire new knowledge in an XR context is a key concern of a scientific approach. In order to allow Cultural Heritage experts to understand the functionalities of living spaces, social activities, as well as interactions, an XR simulation dedicated to heritage must have the capacity to (i) set the historical universe in a functional condition and (ii) put Cultural Heritage experts in situations to evaluate populations’ actions and interactions. These two issues combine consistency of the reconstitution with credibility of the simulation.

48.3

General Methodology

We propose a scientific approach articulated around three cross-functional axes of use: (i) See, (ii) Manipulate, and (iii) Share related to the three challenge families presented above. These three axes correspond respectively to the challenges (i) Perception, (ii) Interaction, and (iii) Collaboration. The three axes are described in the following sections.

1410

R. Gaugne et al.

Fig. 48.3 Overview of the general methodology

All the axes follow the same general methodology starting from Cultural Heritage material and aiming at innovative usages for Cultural Heritage experts. Figure 48.3 presents the different steps of the process. In this methodology, starting from Cultural Heritage that can be artifacts, monuments, excavation sites, or even written documents, paintings, sketches, etc., the first step consists in producing 3D data, either with 3D modeling or by digitization. In this last case, intermediate digital data can be produced, such as photos in photogrammetry, or Dicom data in CT scan, that must be processed to generate 3D data. Many Cultural Heritage projects do not go beyond this step in their usage of 3D data. Even if this can be legitimate in case of illustration or scientific mediation, we prefer to consider these data as an intermediate step on which it is possible to design and product XR content and environments that can be used or evaluated by Cultural Heritage experts. We will present several instances of this methodology to illustrate different context of use or evaluation.

48.3.1 “See” Axis Overview The “See” axis deals with the visualization of both the external and internal structures of archaeological material and its context. The problem here consists in identifying or designing production processes for the data to be visualized as well as modalities for visualizing these data. The production of data relies here on surface digitization technologies such as photogrammetry or laser scanning, as well as on 3D modeling, but also on internal composition detection, mainly using medical image techniques such as CT scan and Magnetic Resonance Imagery (MRI). The viewing modes in an XR environment are based on the properties of the camera that displays the 3D world view. Interacting with this camera allows to navigate the 3D world. In the case of an XR environment, the camera generally corresponds to the point of view of the user, and thus is attached to the position of the user’s head.

48.3.2 “Manipulate” Axis Overview This axis deals with the interaction between the user and the object, whether this interaction is purely virtual or tangible. The goal here is to propose new interaction

48

eXtended Reality for Cultural Heritage

1411

metaphors, based on XR methods and tools. It is important that these interactions are as natural as possible so that the technology fades behind the use and that they integrate into the operational processes of Cultural Heritage experts.

48.3.3 “Share” Axis Overview This axis covers the documentation, sharing, and archiving of digital data. Its objective is twofold: (i) to work on the enrichment of the object at the digital level, by the integration of metadata associated with 3D data and/or the addition of spatialized information within 3D data, and (ii) to allow the researcher to crossreference information on digital corpora. This information can be shared through collaborative or successive work sessions, locally or remotely.

48.4

3D Production

Producing 3D environments with and for Cultural Heritage involves a methodological decomposition that can be adapted to the broad range of issues raised by professionals in this community and the general public. By taking the temporal dimension as the linchpin and scientific knowledge as a driving force, the existing of a Cultural Heritage item can first be digitized, then virtually restored, as it was potentially, at such-and-such time. Figure 48.4 presents the different steps of the process of 3D production, mainly based on two approaches, (i) digitization that relies to the production of 3D data from real material and (ii) graphical design that corresponds to the production of 3D data by 3D modeling.

Fig. 48.4 3D production methodology

1412

R. Gaugne et al.

48.4.1 Digitization and Data Processing In this section, we present several methods to digitize Cultural Heritage material, ranging from simple objects to the complete archaeological site, based on different technologies, with a specific focus on CT scan. Photogrammetry Albrecht Meydenbauer (1834–1921) defined photogrammetry as the science (or technique) that provides reliable information about natural space or physical objects through the recording, measurement, and interpretation of photographs or produced by electromagnetic radiation or other phenomena [16]. More recent and coming directly from photogrammetry, Structure from Motion (SfM) is an imaging technique for estimating 3D structures from 2D image sequences [17]. This technology is still widely used in CH domain thanks to its ease of access and use; see, e.g., [18–22]. 3D cameras, such as the Microsoft Kinect, motion sensing input devices, that can produce depth maps (images in which each pixel contains the distance between the camera and the nearest object) on top of 2D images. Such cameras can produce depth maps by different means (e.g., using Time-of-flight sensors), at different speed and resolutions. Such device provides a great advantage over classical 2D cameras when integrated into a SfM pipeline; see, e.g., [23–25]. An Unmanned Aerial Vehicle (UAV) is an aircraft, without a human pilot aboard, that can collect low-altitude aerial imagery [26–28]. Reflectance Transformation Imaging (RTI) allows to produce a dynamic image from a series of photographs under different artificial lighting [29]. This technology is particularly convenient for the digitization of thin reliefs such as handwriting traces or engraves; see, e.g., [30–32]. 3D scanning is a surveying method performed by a laser scan that allows to obtain 3D coordinates of visible surface points in a fast, contactless, and automatic way. Current laser scanners use different methods of distance measurement [33]. This technology provides precise 3D recording of CH material at various scales [34–38]. Computed tomography (CT) and μ-tomography are X-Ray-based techniques allowing, in particular, the study of the interior of CH items [39]. These techniques make it possible to reconstruct an object volume from a series of measurements made by slices from outside this object. The μ-tomography allows to work on a small scale, ranging from sub-millimeters to a few centimeters, and with a spatial resolution of a few microns. These different techniques are more and more used for a wide range of CH material [40–44].

CT Scan Computed tomography (CT) [45] is an imaging technology widely used in medicine; CT scanners use computer-processed X-rays to produce tomographic images (i.e., virtual “slices”) of specific areas of the scanned object. Data obtained during CT scan consist in a set of continuous attenuation profiles called a sinogram [46]. The sinogram is not directly usable and requires a mathematical processing, a double inverse Fourier transform, to obtain images on which a convolution filter is applied, low-pass filter or high-pass filter, to optimize either

48

eXtended Reality for Cultural Heritage

1413

Fig. 48.5 Data processing with CT scan

density resolution or spatial resolution. The set of images obtained from this tomographic reconstruction (see Fig. 48.5) is used for 2D and 3D post-processing, such as multi-plan reconstruction, 3D reconstruction, and segmentation. Multi-plan reconstruction: this post-processing, also known as 2D MPR, allows to define a new slicing of the initial object in a different plan than the one used during acquisition. 3D reconstruction: this post-processing aims at highlighting, subsequently to the analysis of 2D images, relevant information in a clear and easy to understand way. The highlight of the different components of the material is based on density information calculated during the CT scan, called the radiodensity, and expressed relatively to the Hounsfield scale [47]. The resulting images are also named iconographic images. Segmentation: this post-processing allows to isolate portions of the data in order to focus on relevant parts, and to remove noise data. This technique is particularly important to study nested artifacts.

48.4.2 Graphical Design Distinguished from 3D digitization and more commonly referred to as “graphical design,” 3D modeling of no-longer existing Heritage material leads to a much higher involvement of archaeologists in the sense that their knowledge is the raw material of the restitution process. A large majority of archaeological restitution concerns and has concerned architecture, so we will focus on this kind of reconstitution.

1414

R. Gaugne et al.

Robert Vergnieux synthesizes the heritage uses of 3D architectural restitutions in archaeology [48]. He explains the enthusiasm for them by the immediacy of the visual impact they generate, which greatly facilitates Cultural Heritage enhancement by feeding the collective imagination. Constrained by narrative and playfulness issues, these images, whose photorealism can sometimes freeze information in people’s minds, can nevertheless convey uncertainties, false information, anachronisms, and approximations of details.

Theoretical and Architectural Origins Typology At the beginning of the nineteenth century, Jean-Nicolas-Louis Durand developed an analytical typology based on the plans by creating a catalogue of typological references and rules for architectural design [49]. These concern the built framework and its entire environment. Later other works were carried out on the urban scale [50]. The representation contributes to the typological analysis that brings together the essential properties of a category of real objects and makes it possible to report them sparingly. Representation Modes The “classical” objectives of the representation of built architecture are: • Setting the object of study in its morphology and dimensions • Objectively organizing the relationships between the different elements of the object of study • Getting as full a description as possible with minimal means There are several graphic representation techniques whose choice depends on many factors (size, level of detail, recipient’s profession, practices, agreements, etc.). Among them, we find the geometrical, axonometric, and perspective drawings (cf. Table 48.1).

Architectural Modeling Styles Before discussing the types of restitutions, we can mention two models of archaeological restitution developed in the nineteenth century: • Ruskin represents a “conservationist” point of view and advocates “non-intervention” on the remains [52]. • Viollet-le-Duc advocates the total reconstruction of the ruins: To restore a building is not to maintain, repair, or rebuild it, it is to restore it to a complete state that may never have existed at a given time [53].

48

eXtended Reality for Cultural Heritage

1415

These two trends have led to the existence of six modern restitution models presented in Table 48.2. We can mention three types of geometric 3D models with an architectural vocation and based on surveys [54, 55]: • As surveyed: It is the object as it exists at the time of the survey and its most accurate modeling. Table 48.1 Techniques for the graphical representation of built architecture [51]

Projection

Representation of the subject

Geometrical drawing: plans, sections, and elevations Orthogonal on at least two planes As it is in itself without distortion of perspective

Axonometric drawing Orthogonal on one plane

As it would be seen by an observer placed in infinity

Perspective drawing Conic from a given point of view on a plane As the observer sees it in reality

Table 48.2 Modern restitution models Restitution model Restitution of the monument on site

Restitution of volumes Transfer model

Replica model

Conservationist model Virtual model

Definition Reconstitute the architecture or monuments on the remains, using the materials and techniques that have been documented on the site, in order to give the buildings an appearance similar to that they had at the time of the site’s life Action executed on the remains, but using materials clearly different from those used on the site Moving a site or, more frequently, a part of the site (one or more buildings) to a different location from the original one Partial or complete restitution of a site on a separate space – even far – from the site, i.e., on a space where there are no archaeological remains Minimal intervention, which is normally limited to consolidation or restoration of the remains Show the restitution of the site (or part of the site) without physically materializing it, normally using graphic means

Examples Pompeii, Herculaneum, Martigues (Bouches-du-Rhône, France), and Augusta Raurica (Switzerland)

Piazza Armerina (Sicily) and Xanten Thermae (Germany) Skansen on the island of Djurgården (Stockholm), Kolomenskoye Park (Moscow), and Latenium (Neuchâtel, Switzerland) Plymouth (Massachusetts)

Consolidation in most archaeological sites in Europe No-longer existing sites such as the first Palais de l’Intendant in Québec, or the Roman Theater in Fano, Italy

1416

R. Gaugne et al.

• As built: It is the object without any extrapolation. The model is used to mask the effect of stone erosion and minor defects due to time. The objective is to get the original state of the element detected. • Reconstitution: The measurements are completed by the expertise of the specialist: archaeologists, architects, etc. They validate all the assumptions made during the modeling process to ensure that the model is complete. The fundamental principle of any restitution is the guarantee of the conservation of the site or the remains [56]. There are several fields concerned by the restitution procedures: the scientific field of archaeology, education, tourism, and heritage presentation. In the following, we will only deal with aspects of production and scientific mediation. Scientific production Whatever its model, the scientific interest of the restitution allows the verification of the hypotheses made about the old buildings [57]. For the virtual model, Robert Vergnieux indicated in 2011 that in the scientific context, the restitution process is complex, multidisciplinary, and requires years of research. The methodological objective of 3D models is to be able to raise all the validation questions involved in the restitution work. Unlike 3D models for illustration purposes, the idea is to include all the structural units of the buildings that make it up. The reader of the image must also be able to identify the uncertainties, the chronological phase(s) selected, the version, and the documentary sources. More recently, if respect and management of documentary sources seem to be applied [58], the way in which constraints, choices and points of view of restitution influence and alter the relationship to data and sources, raises questions [59].

48.5

XR Design and Implementation for Cultural Heritage

There are many technological possibilities to implement an XR application offering different experiences and interactions for the user [60–62]. In this section, we will present the main different systems and interactions that can be proposed, and we will develop two notions of interaction that can be used in a Cultural Heritage context. Figure 48.6 presents the different steps of the process of XR design and implementation in this context.

48.5.1 XR Systems for Cultural Heritage In this section we briefly present the common technologies of XR. Since each type of XR system has its own characteristics and some might prove more suitable for certain use cases of CH, we present a clear taxonomy of VR/AR technologies and present their advantages and limitations.

48

eXtended Reality for Cultural Heritage

1417

Fig. 48.6 Overview of the general methodology for XR design and implementation

Virtual Reality Systems CAVE

Workbench

HMDs

[63] are high-end VR facilities comprising of multiple walls on which stereoscopic 3D content can be projected in real-time. Users can interact at scale 1 with virtual objects. Such systems are generally single-user even if some seldom facilities propose multiple-users implementation. But in this last case, the system raises occlusion issues between the user and the virtual scene. CAVE-like systems are very comfortable for the user as he/she is present with his/her own body in the simulation and can move naturally within the physical space [64–66]. The main drawback of this kind of system is its cost This term is usually used to refer to a combination of one or more screens used to display stereoscopic 3D content. Unlike Caves, workbenches do not necessarily allow for a scale 1 interaction with virtual objects. However, their limited size and cost make them an interesting system to provide interactions in a stereoscopic environment. In particular, they can represent an efficient tool to work on limited size objects such as the ones studied in archaeological laboratories [67] Head-Mounted Displays are headsets that can be used to display 3D stereoscopic content. Users are occluded from the real world and are thus more immersed than within a Cave or with a workbench. The main drawback of this kind of system is that the user loses the relation with his/her own body, but HMDs have become very popular thanks to their low price. They are widely used in Museum to provide immersive experiences for the visitors

Augmented Reality Systems VST

Video see-through [1] AR are either “Hand Held” systems based on tablets or smartphones or “Head Mounted Displays” with cameras such as Windows Mixed Reality headsets. In this case, the real scene is filmed through the cameras of the device, and the digital content is then inserted to be displayed on the screen (or screens) of the VST (continued)

1418

OST

SAR

R. Gaugne et al.

device. This kind of AR systems is very popular in Museum contexts as they just require a smartphone or a simple tablet. Interactions and complexity of simulations proposed with VST are limited by the computing performances of the device, and the display is monoscopic. HMDs provide stereoscopic rendering, but the quality of the view of the real world is dependent of the performance of the cameras of the device Optical see-through [1] AR are systems where the user sees reality directly through semitransparent screens in front of his/her eyes, which insert some digital content, like the Microsoft HoloLens and the Magic Leap. This kind of system is increasingly used as it provides a natural view of the real world with hand tracking for interactions. The current versions of these devices are still perfectible in terms of field of view and tracking, but the applications in CH are promising [68] Spatial (or Spatially) Augmented Reality [69] is a term used in the AR community to characterize projective AR where 3D digital content is directly projected on the reality, i.e., onto real objects. This kind of system is widely used in cultural performances on monuments but can also be used to project surface textures on Cultural Heritage objects such as engraves or statues with particular calibration and deformation processes [70, 71]. This kind of system can also be used to display the internal content of an object, but in this case the system must deal with depth perception and integrate a head tracking to display the correct user’s point view [72]

48.5.2 XR Interaction and Collaboration in Cultural Heritage In this section, we present general interaction concepts in XR, and the main kinds of interaction according to [73]. • Navigation: This interaction with user’s point of view is fundamental as it is the main and often only interaction with the virtual universe. In this case, the user interacts with the camera related to his/her point of view. • Interaction with virtual environment: This family of interactions is related to the modification of the characteristics of some objects that belong to the simulation of the virtual environment. We detail below this family of interaction that opens many interesting possibilities for Cultural Heritage domain to benefit from XR. • Interaction with application: This family of interactions is related to the modification of the properties of the XR application through dialog objects that are not related to the simulation. An interaction is defined by several characteristics. The first concerns the modality(ies), i.e., the means of the users that are implemented to effect the interaction (hand, body, gaze, brain). A second characteristic is the metaphor, i.e., a representation of the interaction in the virtual world to make it understandable to the user (virtual hand, pointer, radius, virtual tool). The last characteristic is the device used, i.e., a technical device used in the interaction. These different characteristics are widely documented in the literature to identify the most adapted and relevant ones according to a specific domain and specific task.

48

eXtended Reality for Cultural Heritage

1419

Fig. 48.7 Operational interactions with engraves in Augmented Reality (left) and Virtual Reality (right)

In Cultural Heritage applications, we propose to distinguish between two families of interactions with the virtual environment, (i) functional interactions that put in function objects, monuments, or sites and (ii) technical interactions, such as annotations, cutting plans, or taking measurements. Functional interactions are associated to a simulation of the functioning of an archaeological context, be it a site, a building, or an object. They allow the user to put these artifacts into operation in the simulation. In the reconstitution of industrial heritage presented in [74], the user can interact with complex elements of an old power plant. In the example of the reconstitution of the eighteenth-century East India Company, Boullongne [75], the user can manipulate the ship’s steering wheel and vary the navigation heading, or ring the quarter change bell. This kind of interaction is not unique to archaeology. In the medical field, the user can interact with surgical instruments and perform surgery [76–78], or in the industrial field the user can interact with complex machine tools simulated in Virtual Reality [79–81]. Operational interactions can act on the representation of an object, a building, or a site to generate study data or new representations as part of a business process. Thus, in [71], using a projective augmented reality system, stone etchings are augmented by a numerical representation in order to highlight the relief (Fig. 48.7 on the left). In a second example presented in [82], engravings in a room in Barnenez’s cairn are highlighted in a virtual reality reconstruction, based on photogrammetric scanning (Fig. 48.7 on the right).

Functional Interactions In the following, we present examples of functional interactions implemented in the framework of cases of archaeological studies carried out during different projects. These examples are intended to illustrate these cases of use of interactions in order to stimulate reflection on its interest in the archaeological operating chains. Functional Interaction with an Object: Osse´’s Weight The granite weight of Ossé (La Claraiserie, Ille et Vilaine, France) was found in a farm dating from the second or first century before our era, in a deposit of metal

1420

R. Gaugne et al.

Fig. 48.8 Actual weight and steelyard (top, left), digitization and modeling (top, center, and right), and functional interaction in historical context, in Virtual Reality (bottom)

objects, as part of preventive excavations conducted by Inrap, the French National Institute for Preventive Archaeological Research. To better understand the use of this weight, we proposed in [12] a modeling of a roman steelyard, based on fragments found on the same site, and on a complete balance found on another similar site, and a reconstruction of the physical behavior, in a virtual environment representing a site of the same era; see Fig. 48.8. The user immersed in the virtual universe can then weigh objects and put back into operation this scale: he can catch bags with different sizes and hang them to the hook, and move the counterweight to balance the system and measure the weight of the bag. In addition, we used a copy of 3D print weight as a tangible object for interacting in virtual reality simulation. Thanks to an infra-red camera position capture system, this object is recognized and tracked in the environment, and the user can use this weight in interaction with the scale by weighing virtual objects. Functional Interaction with a Building: The Boullongne Boullongne, ship of the East India Company of the eighteenth century, was the subject of a special study by the department of Maritime History of the University South Brittany, because of the conservation of numerous documents such as logbooks, and naval architecture technical plans. To allow a better understanding of life on board, it was decided to build a virtual replica of this ship. The modeling was carried out by the West Digital Conservatory for Archaeological Heritage project [83] from the plans and a model of the ship. From this modeling, we put this vessel into operation by a physical simulation of the vessel and its environment as presented in [75]. We first worked on the waterline of the boat on a simulation of the water so that the ship moves according to the state of the sea. In a second step, we proposed interactions to the user so that the user can interact with the objects constituting the

48

eXtended Reality for Cultural Heritage

1421

Fig. 48.9 Physical simulation of the Boullongne (top) and functional interactions with the wheel and the bell (bottom)

ship. For example, it is possible to turn the bar of the ship with a physical behavior that transforms the angle of the bar into forces that act on the boat and thus modifies its trajectory. We also added the possibility of ringing a bell or firing a cannon. Figure 48.9 illustrates the physical simulation of the ship and the functional interactions with the wheel and the bell. Functional Interaction with a Site: Re´mire The Rémire site is a former sugar plantation run by the Jesuits in the seventeenth and eighteenth centuries in Loyola, French Guiana. For this site, studied by a team of archaeologists from Université Laval in Quebec, we endeavored to simulate its link with the natural environment [84]. The different buildings of the site and its topology were modeled by the West Digital Conservatory for Archaeological Heritage project [83]. Particular attention was paid to the reconstitution of the vegetation of the site, with the modeling of representative trees and plants. We have reconstructed the race of the sun, the moon, and the celestial vault, for a day of the studied time, with the representation of the shades and lighting (Fig. 48.10, center). The user immersed in the environment can change the time scale to easily observe different times of the day (Fig. 48.10, bottom). We also simulated, according to the time of day, the sound environment consisting of animal songs (frogs, birds) by zone according to the plant environment (Fig. 48.10, top, right).

Operational Interactions In the following, we present examples of operational interactions implemented in the context of case studies that we have carried out during different projects with archaeologists. Again, these examples are intended to illustrate this case of use of interactions to feed the reflection on its interest in the archaeological operating chains.

1422

R. Gaugne et al.

Fig. 48.10 3D model of the site (top, left), sound simulation of the natural environment of the site (top, right), simulation of the night and day course (center), and functional interaction in Virtual Reality (bottom)

Virtual Excavation: The Urn of Guipry Archaeological excavation, whether carried out on a site scale or on an archaeological material scale in laboratory (Fig. 48.11, top, left), is a destructive activity for which there is no backtracking or possible replay. Moreover, a search is most often done with no visualization and limited or no knowledge of the underlying content that will be found. Finally, some elements cannot be detected, and their traces are destroyed during the search. We propose augmented and virtual search environments which allow, upstream, to better prepare this real search, in complement, to accompany the real excavation, to document it, or a posteriori, to return on certain elements of the site of excavation. We evaluated in [68] a Mixed Reality system to help preparatory archaeological process (Fig. 48.11, top, right), and we illustrated, in [67], the use of a Virtual Reality tool for the search of a funerary urn of Guipry (in Ille et Vilaine, France). In this example, the user can manipulate, remove the objects from the ballot box, measure them, or annotate them. He can also manipulate cutting plans to visualize the internal structure of the urn. This application, here applied to the Guipry urn, is generic in its design; it can be applied to any 3D model and offer the same features and interactions. The first implementation of the application was deployed on a workbench virtual reality environment (Fig. 48.11, bottom) that was best suited to the size of the archaeological material being studied.

48

eXtended Reality for Cultural Heritage

1423

Fig. 48.11 Actual micro-excavation (top, left), Mixed Reality system to visualize inside the urn (top, right), Virtual Reality system to perform virtual excavation (bottom)

Point Cloud Manipulation: The Real Tennis Building of Rennes A point cloud is the raw data obtained during laser scanner scanning. This data is generally transformed into a surface mesh to be exploited in 3D. However, the point cloud constitutes in itself 3D data that can be exploited in XR, and which constitutes the most faithful data, the surface generation processes being able to induce errors. This representation is little used so far, because it is complex to implement in an interactive application, especially because it does not benefit from graphic optimization techniques widely developed in computer graphics on the rendering of polygons. We have implemented a number of tools that allow to interact directly with the point cloud, in VR (Fig. 48.12), including an automatic pipeline to transform the PLY point cloud into an octree-structured billboard based on the principle of [85]. For this case study, we are interested in the Real Tennis court of Rennes [86]. The scan of the building generated of point cloud with more than 780 million points. The 3D model generated from the initial point cloud is constituted of more than 500 millions of triangles which is realtime rendered using LOD and occlusion culling optimizations. We offered the user several interactions and tools such as 1:1 navigation, scaling, density change, LOD, measurement, cutting plane, 2D plane, and integration of photo, sketch, and 3D model views [87], as illustrated in Fig. 48.12, bottom. These interactions were designed in collaboration with the archaeologists in charge of the excavation of the building in order to help them in their study: as they had a short time access to the real building, they wanted to be able to come back later into the virtual building, to perform some measurements, to review architectural details, and to document reports.

1424

R. Gaugne et al.

Fig. 48.12 Immersive visualization of the point cloud at 1:1 (top, left), example of 2D cut view obtained with the tool (top, right), interactions with the point cloud in high definition (bottom, left), cutting plan manipulation with a tracked tactile tablet (bottom, center), and position of a photographic view in the Virtual Reality application (bottom, right)

48.6

Evaluation and Use by Cultural Heritage Experts

The methodology developed in this chapter aims at proposing new processes and new tools for Cultural Heritage experts. It is thus important to allow these experts that can be archaeologists, historians, or curators, to evaluate them on various use cases. We chose to apply our approach to several actual use cases in order to validate the usefulness of the tools and the efficiency of the processes based on XR applications as presented in Fig. 48.13. In this section, we detail different use cases and discuss their contribution to Cultural Heritage domain.

48.6.1 Study of a Cremation Urn The cremation urn presented in section “Operational Interactions” was studied following the workflow presented in Fig. 48.14. The urn was digitized using CT scan generating Dicom data. This data was processed using segmentation and 3D surface generation based on radiodensity, in the open source software Horos. This allowed to generate 3D models for the internal content of the urn, especially bones fragments and metal parts. The segmentation and 3D model generation of a fibula inside the urn was 3D printed and studied by archaeologists before the physical excavation of the urn [88]. The internal content digitization was also used to produce a complex transparent 3D printing of the whole urn and its content [89]. We designed and implemented two XR applications in order to help archaeologists in the study of such material as presented in section “Operational Interactions.”

48

eXtended Reality for Cultural Heritage

1425

Fig. 48.13 The evaluation and use of XR applications through use cases

Fig. 48.14 Case study of the cremation urn of Guipry

The digitization process based on CT scan appeared to be very well adapted to the context of such archaeological material. Volume reconstructions from CT scan allow spatial localization of all artifacts and their identification. If modeling indicates how the objects were organized in the burial (in particular the cremated bone), it mainly serves to rapidly determine if the deposits contain sensitive objects such as metal. This information allows to anticipate optimally fine analysis of clusters and to implement protective measures for artifacts. 3D acquisition associated to 3D printing of particular internal element such as the fibula offers the possibility to duplicate an artifact among others, leading to an immediate typological identification of the artifact, as well as morphological, typometric, or technological observations, regardless of contingencies relative to conservation and restoration operations often required by such objects. This process allows the provision of information within

1426

R. Gaugne et al.

hours, which is not the case for the “operational chain” commonly implemented that can take months if protective measures are implemented, prior to the study of the archaeological material. The complex transparent provides a 3D print of the container and contents (here funeral deposit burned bones and iron objects, knife rivet, and fibula). Its tangible nature allows “direct” access to information and the physical handling of cremation deposit. It allows viewing and manipulation of the vessel and its contents in full and with integrity. Printing allows visualization of burned bones that are imperceptible when excavating (too small, fragile, or appearing negative, to be taken), but also some well-preserved bones (at least here, a long bone) and the determination of iron artifacts (here released from their mineralization). 3D printing also provides information for the reconstruction of the funerary urn. This tangible medium allows a manipulation/simple visualization to work on the analysis (such as the distribution of artifacts and the burial gestures) but also as a support for the excavation. Indeed, 3D printing is the only tangible medium of context pre-served after the excavation of the incineration. The mixed reality system associates in the same space, the real (cinerary urn) and 3D models obtained from reconstruction. This possibility is perceived by the archaeologist user as a revolution. The system was presented to archaeologists expert in the study of such material. They confirmed the good spatial perception and were able to precisely identify eight different plans of depth, i.e., to differentiate up to eight elements one behind the other, within the urn, to correctly locate the different artifacts, and evaluate the distance between them. The virtual reality system INSIDE was evaluated by archaeologists to identify the advantages and drawbacks of the method. The virtual introspection interactive tool was confirmed to be really useful to get more information, as the visualization inside the urns shows well the different dispositions of the content, with 3D fragments at the real scale. The identification of fragments was eased by the manipulation of each fragment on its own, to isolate them from the rest. Contrary to the direct visualization of the CT scan, which requires some expertise to get the correct points of view, virtual reality provides natural 3D interaction such as manipulating a cutting plane. Even though the information displayed is limited by what the scan could give (density and form), it is enough relevant to prepare a real excavation, to concentrate the effort on the regions of interest.

48.6.2 Reconstitution of an Eighteenth-Century Ship The eighteenth-century ship, the Boullongne, introduced in section “Functional Interactions” was reconstituted based on Historical documentary material as presented in Fig. 48.15. Original architectural plans were used to 3D model the ship at 1: 1. The model was integrated in a maritime simulation, with several interactions in order to simulate the functions of the ship. The goal of this work was to enable Historians to be immersed in the subject they were studying and to better understand

48

eXtended Reality for Cultural Heritage

1427

Fig. 48.15 Case study of the eighteenth-century ship, the Boullongne (Source: Musée de la Compagnie des Indes – Ville de Lorient)

the life environment of eighteenth-century sailors in the French East India Company that was testified in archived logbooks of this period. The virtual reconstitution of La Compagnie des Indes Orientales ship allows a new practice of validation of the historical sources in an innovative application environment, involving historians’ own physical activities and their perceptions. Being positioned on a ship in scale-one enables maritime historians to appreciate living and working conditions in a small space, overpopulated and in perpetual movement. The historical reconstitution of Le Boullongne allows historians to understand the architecture of a merchant ship and especially to assess the dimensions and volumes of the different spaces, from the hold to the deck. Historians try to figure out how life could be organized on the ship during trips that lasted between 5 and 6 months, with maximum congestion and strong promiscuity, as the ship carried between 130 and 150 persons. For example, the transition from forward to stern on the upper deck, or the height of the steerage where the guns are placed, seemed extremely narrow to historians. This raises the question of the actual use of guns, and the concrete movement of sailors and passengers on the upper deck. All these thoughts and discussions between historians were aroused following the natural 1:1 exploration of the ship. In particular, the necessity of walking in a crouching stance in the intermediate decks evoked in Historians a study of historical archives of La Compagnie des Indes Orientales which highlighted a particular

1428

R. Gaugne et al.

pathology for sailors: most of them suffered a hernia. Carrying and storing goods in the confined cargo holds, with no possibility of standing up, may partly explain this pathology. A specific work dedicated to the study of human activities in the ship used motion capture and graphical design of Historically consistent characters to reconstitute some life scenes. The real activities were designed and played by Historians, and recorded with a Mocap optical system. The scenes were integrated as triggered animations inside the VR application [90]. In this particular work, instead of usual objects or monuments, a particular human activity is digitized, relying on intangible heritage. Such simulation allows historians to sail on a disappeared ship. It enables a better understanding of how on-board life was organized, providing an immersive selfexperience of the behavior of the ship. It also constitutes a powerful pedagogical tool, giving life to a valuable testimony of our history. The virtual sailing was presented to several public exhibitions.

48.6.3 Process of Analysis of a Gallic Grave In the context of a preventive archaeological investigation, an exceptional aristocratic Gallic grave at Warcq was discovered and excavated by Inrap. The remains discovered in the grave have revealed to be exceptional. The vast funerary chamber (5.50  2.80 m) was preserved to a depth of more than a meter. In this damp environment, the wooden walls and ceiling were extremely well preserved. Over time, the latter had collapsed directly onto the floor of the chamber, covering the deceased and his goods. The funerary artifacts discovered were of unusually high quality. The main item was a two-wheeled ceremonial chariot. One of the most spectacular elements was the burial of four horses around the vehicle. The deceased person, most likely a man, was lying on the body of the chariot. As the ground environment of the excavation in Warcq was very wet, it was decided to sample complete blocks in the grave around some pieces of interest that required a cautious excavation process [91]. Seven blocks were scanned in order to prepare their micro-excavation and restoration. The block S1 contains the wheel hub, the blocks S2 and S3 contain two buckets, the block S4 contains the rim of one wheel, the blocks S5 and S6 contain the harnessed heads of the two horses ahead the chariot (Fig. 48.5), and the block S7 contains a yoke. The data generated during the CT scan was processed in order to perform preliminary studies and segmentation and generate volume rendering and surface rendering. CT scan data (sections, 3D reconstruction) guided all stages of microexcavation. These operations were carried out in collaboration between the archaeologist and the curator. The whole process of micro digging has also been documented using photogrammetry. All the samples followed the same process: the micro-search, guided by CT data, under a magnifying glass, is followed by a cleaning for the study by micro-sandblasting under a binocular loupe.

48

eXtended Reality for Cultural Heritage

1429

This process has been particularly efficient for the study of two harnessed horse heads where it has allowed the researchers to observe, quantify, and locate the presence of objects inside the samples upstream of the intervention, despite the poor state of conservation. The information was very useful to assist the restorer during micro-excavation. The most important example is the decoration of the harness. These visible elements on the images from the CT scan are very small. They were composed of very thin sheet of copper alloy originally covering the nail heads. They were not all found during the micro digging because of their size, the very complex sediment to search through, and their state of conservation. Imaging and direct observations on the micro frame made it possible to restore the different elements of the harness (Fig. 48.16). We used 3D surface rendering to generate 3D meshes corresponding to different density ranges in order to produce a 3D print that highlights the spatial organization of the metal parts of the harness. We chose to use a technique of transparent printing of the block, with different colors for the internal elements. The transparent 3D printing constitutes a tangible spatial recording of the horse head, with the nails visible; it offers a better visualization and perception of the volume of spatialization and of the relation between the objects. The head itself was not restorable aside from the teeth, as there is no more bone matter. 3D printing has thus a long-term goal of safeguarding. The digital data of the different samples were integrated in a full restitution of several steps of the excavation of the complete grave, in virtual reality. Both photogrammetric and CT data of samples were integrated in the reconstitution and manipulable by the user with the possibility to switch between the different views. Several tools of operational interactions such as measurement, magnification, and

Fig. 48.16 Case study of the aristocratic Gallic grave of Warcq

1430

R. Gaugne et al.

X Rays view of the objects are proposed to the user. Archaeologists in charge of the excavation tried the virtual reconstitution. The first comment was that they never had a complete view of the grave during the excavation because of the presence of scaffolding. This complete view gave them a new perspective of the organization of the grave. In this work, computed tomography and 3D printing are tools within an “operating chain” of analysis that involves the interaction of different experts. Virtual reality provides different views of the site, with the access to different steps of the excavation and a complete overview of the grave. These views were impossible in the actual grave; they open new perspectives to apprehend a site and its excavation process.

48.6.4 Annotation of Megalithic Art with Augmented Reality Megalithic art is a spectacular form of symbolic representation found on prehistoric monuments. Carved by Europe’s first farmers, this type of art allows an insight into the creativity and vision of prehistoric communities. As examples of this art continue to fade, it is increasingly important to document and study these symbols. In [14] we introduced MAAP Annotate, a Mixed Reality annotation tool developed as part of a project (the Megalithic Art Analysis Project – MAAP) with the school of Archaeology of the University College Dublin (UCD), Ireland. It provides an innovative method of interacting with megalithic art, combining cross-disciplinary research in digital heritage, 3D scanning and imaging, as well as Augmented Reality (AR); see Fig. 48.17.

Fig. 48.17 Megalithic Art annotation with Augmented Reality

48

eXtended Reality for Cultural Heritage

1431

We put a particular attention in the scanning process, with a portable autonomous system based on the use of a laptop and a Microsoft Kinect to offer archaeologists the possibility to scan in real-time the art directly onsite (i.e., in an outdoor environment). Once the art was scanned and a 3D model was obtained, we focused on the development of an AR annotation tool for megalithic art. It was designed to be used by archaeologists and not by computer scientists as a unique mean to examine megalithic art offsite. Indeed, one of the issues with megalithic art consists of being able to determine what is really art and what is only due to the natural erosion of the stones. MAAP Annotate was thus designed with input from archaeologists who had no prior experience with Mixed Reality devices in order to provide them with the tools they need in achieving this distinction between art and erosion. We thus offered them the possibility to manipulate the scanned 3D stones (i.e., to change their scale, to rotate them, etc.) as well as natural selection tools to identify art on the stones and more user-specific tools such as a virtual flashlight that is used to cast shadows on the 3D stone to better detect shallow engravings. An evaluation of the resulting tool was carried out on a group of 10 experts (archaeologists from UCD’s school of Archaeology), using an optical see-through HMD. The test confirmed that this application is usable and that archaeologists view the use of Augmented Reality as an excellent way to annotate megalithic art. One interesting feedback about this tool is the possibility of being able to view a stone and take notes simultaneously without being interrupted by the necessity to write on paper. This characteristic of systems based on a Mixed Reality HMD device was also pointed out and appreciated in the tool presented in Sect. 48.6.1.

48.7

Conclusion

Methods and tools for Cultural Heritage based on eXtended Reality open new perspectives and research questions but also new ways of understanding the professions of Cultural Heritage practitioners. Interactions with digital objects in eXtended Reality allow Cultural Heritage experts to work in a non-destructive way on archaeological or historical material that is generally extremely fragile, often difficult to access, and sometimes not visible to the naked eye. 3D interaction tools provide support to the excavation or restoration operations for the archaeologist. For example, it is possible to prepare an upstream operation, or guide it during its realization. In this case, the proposed interactions are said to be operational. In addition, 3D data can be integrated into a richer virtual world that allows one to contextualize the artifact, be it a context of manufacture, use, or discovery. In this case, it is possible to propose interactions with a simulation of the function of the artifact. In this case, the proposed interactions are said to be functional. These tools have shown to also provide support to historians by allowing them to be immersed in historical contexts that no more exist.

1432

R. Gaugne et al.

This research also represents new scientific challenges in the realm of virtual, augmented, or mixed realities. The perception of 3D universes through the specificity of the archaeologist as a user is a very promising area. As a professional expert, the Cultural Heritage expert does not necessarily look for a photorealistic representation of his study space, like the doctor who studies his patient through an X-ray or MRI. The scale-1 interaction with physical displacement is in itself an interesting 3D digital world approach in archaeology and more generally Cultural Heritage, as it combines a mode of perception, proprioception, and a fundamental interaction, natural navigation, which are not accessible through a simple screen keyboardmouse. Interactions with multi-scale areal or surface objects represent challenges that remain to be overcome and that will allow access to data of great richness. The interaction with tangible objects in mixed reality opens an important field of investigation, especially for interpretation in multiple contexts. Finally, the Cultural Heritage expert can dynamically interact on several levels: he seeks the links between the elements of a site, his own interactions with these elements, but also the interactions that could have humans at the time with this site. It is therefore also important to propose representations and simulations of human activity in contextualized environments.

References 1. Milgram P, Kishino F (1994) A taxonomy of mixed reality visual displays. IEICE Trans Inf Syst E77-D(12):1321–1329 2. Fuchs P, Moreau G, Guitton P (2011) Virtual reality: concepts and technologies, 1st edn. CRC Press, Boca Raton 3. Pimentel K, Teixeira K, Jamet D (1994) La réalité virtuelle. . .: de l’autre côté du miroir. Ed. Addison-Wesley 4. Azuma RT, Baillot Y, Behringer R, Feiner S, Julier S, MacIntyre B (2001) Recent advances in augmented reality. IEEE Comput Graph Appl 21(6):34–47 5. Reilly P (1990) Towards a virtual archaeology. In: Computer applications in archaeology. British Archaeological Reports, Oxford, pp 133–139 6. Krasniewicz L (2000) Immersive imaging technologies for archaeological research. In: CAA: the 28th conference on computer applications and quantitative methods in archaeology, vol 843, pp 163–169 7. Pujol-Tost L (2011) Realism in virtual reality applications for Cultural Heritage. Int J Virtual Real 10(3):41 8. Forte M (2011) Cyber-archaeology: notes on the simulation of the past. Virtual Archaeol Rev 2(4):7–18 9. Christou C, Angus C, Loscos C, Dettori A, Roussou M (2006) A versatile large-scale multimodal VR system for Cultural Heritage visualization. In: Proceedings of the ACM symposium on virtual reality software and technology, VRST ‘06. ACM, New York, pp 133–140 10. Vote E, Feliz DA, Laidlaw DH, Joukowsky MS (2002) Discovering Petra: archaeological analysis in VR. IEEE Comput Graph Appl 22(5):38–50; Acevedo D, Vote E, Laidlaw D, Joukowsky M (2001) Archaeological data visualization in VR: analysis of lamp finds at the Great Temple of Petra, a case study. https://doi.org/10.1109/VISUAL.2001.964560 11. Forte M, Kurillo G (2010) Cyberarchaeology: experimenting with teleimmersive archaeology. In: 2010 16th international conference on virtual systems and multimedia, pp 155–162

48

eXtended Reality for Cultural Heritage

1433

12. Nicolas T, Gaugne R, Tavernier C, Petit Q, Gouranton V, Arnaldi B (2015) Touching and interacting with inaccessible Cultural Heritage. Presence Teleop Virt 24(3):265–277 13. Lim C-K, Cani M-P, Galvane Q, Pettré J, Zawawi TA (2013) Simulation of past life: controlling agent behaviors from the interactions between ethnic groups. In: 2013 digital heritage international congress (DigitalHeritage), vol 1, pp 589–596 14. Barbier J, Kenny P, Young J, Normand J-M, Keane MT, O’Sullivan M, Ventresque A (2017) MAAP annotate: when archaeology meets augmented reality for annotation of megalithic art. In: 23rd international conference on virtual system & multimedia, VSMM 2017, pp 1–8 15. Hermon S, Nikodem J 3D Modelling as a scientific research tool in archaeology. In: Posluschny A, Lambers K, Herzog I (eds) Proceedings of the 35th international conference on computer applications and quantitative methods in archaeology, Apr 2007, pp 140–145 + CD–ROM 16. Slama CC (1980) Manual of photogrammetry. Technical report, America Society of Photogrammetry 17. Ullman S (1979) The interpretation of structure from motion. Proc R Soc Lond B 203(1153): 405–426 18. Aicardi I, Chiabrando F, Lingua AM, Noardo F (2018) Recent trends in Cultural Heritage 3D survey: the photogrammetric computer vision approach. J Cult Herit 32:257–266 19. Dhanda A, Reina Ortiz M, Weigert A, Paladini A, Min A, Gyi M, Su S, Fai S, Santana Quintero M (2019) Recreating Cultural Heritage environments for VR using photogrammetry. Int Arch Photogramm Remote Sens Spatial Inf Sci XLII-2/W9:305–310. https://doi.org/10.5194/isprsarchives-XLII-2-W9-305-2019 20. Anderson EF, John D, Mikulski R, Redford A, Romero M (2020) Preserving and presenting Cultural Heritage using off-the-shelf software. In: Visual computing for Cultural Heritage. Springer, pp 423–444 21. Pierdicca R, Frontoni E, Malinverni ES, Colosi F, Orazi R (2016) Virtual reconstruction of archaeological heritage using a combination of photogrammetric techniques: Huaca Arco Iris, Chan Chan, Peru. Digital Appl Archaeol Cult Herit 3(3):80–90 22. Quattrini R, Nespeca R, Ruggeri L (2017) Digital photogrammetry for archaeological artefacts acquisition. In: IMEKO international conference on metrology for archaeology and Cultural Heritage, pp 643–648 23. Smisek J, Jancosek M, Pajdla T (2013) 3D with kinect. In: Consumer depth cameras for computer vision. Springer, pp 3–25 24. Cappelletto E, Zanuttigh P, Cortelazzo GM (2016) 3d scanning of Cultural Heritage with consumer depth cameras. Multimed Tools Appl 75(7):3631–3654 25. Allegra D, Gallo G, Inzerillo L, Lombardo M, Milotta FLM, Santagati C, Stanco F (2016) Low cost handheld 3d scanning for architectural elements acquisition. In: Eurographics Italian chapter conference, pp 127–131. 26. Mancini F, Dubbini M, Gattelli M, Stecchi F, Fabbri S, Gabbianelli G (2013) Using unmanned aerial vehicles (UAV) for high-resolution reconstruction of topography: the structure from motion approach on coastal environments. Remote Sens 5(12):6880–6898 27. Jo YH, Hong S (2019) Three-dimensional digital documentation of Cultural Heritage site based on the convergence of terrestrial laser scanning and unmanned aerial vehicle photogrammetry. ISPRS Int J Geo Inf 8(2):53 28. Themistocleous K, Mettas C, Evagorou E, Hadjimitsis D (2019) The use of UAVs and photogrammetry for the documentation of Cultural Heritage monuments: the case study of the churches in Cyprus. In: Earth resources and environmental remote sensing/GIS applications X, vol 11156. International Society for Optics and Photonics, p 111560I 29. Mudge M, Malzbender T, Schroer C, Lum M (2006) New reflection transformation imaging methods for rock art and multiple-viewpoint display. In: VAST, vol 6, pp 195–202. Citeseer 30. Wei W, Huang L, Zhu X, Ling L, Guo K, Huang H (2019) Application of reflectance transformation imaging for the display of handwriting traces. Chin Opt Lett 17(11):111101 31. Saunders D, Collmann R, Siddall L (2016) Reflectance transformation imaging and the cuneiform in Australia and New Zealand collections project. Buried History 52:45–48

1434

R. Gaugne et al.

32. Gaugne R, Labaune F, Fontaine D, Le Cloirec G, Gouranton V (2020) From the engraved tablet to the digital tablet, history of a fifteenth century music score. ACM J Comput Cult Herit, to appear 17:1–18. https://doi.org/10.1145/3383782 33. Boehler W, Heinz G, Marbs A (2002) The potential of noncontact close range laser scanners for Cultural Heritage recording. Int Arch Photogramm Remote Sens Spat Inf Sci 34(5/C7):430–436 34. Kushwaha SKP, Dayal KR, Raghavendra S, Pande H, Tiwari PS, Agrawal S, Srivastava SK et al (2020) 3D digital documentation of a Cultural Heritage site using terrestrial laser scanner – a case study. In: Applications of geomatics in civil engineering. Springer, pp 49–58 35. Quagliarini E, Clini P, Ripanti M (2017) Fast, low cost and safe methodology for the assessment of the state of conservation of historical buildings from 3D laser scanning: the case study of Santa Maria in Portonovo (Italy). J Cult Herit 24:175–183 36. Fantoni R, Almaviva S, Caneve L, Caponero M, Colao F, Ferri De Collibus M, Fiorani L, Fornetti G, Francucci M, Guarneri M et al (2017) Laser scanners for remote diagnostic and virtual fruition of Cultural Heritage. Opt Quant Electron 49(3):120 37. Lezzerini M, Antonelli F, Columbu S, Gadducci R, Marradi A, Miriello D, Parodi L, Secchiari L, Lazzeri A (2016) Cultural Heritage documentation and conservation: threedimensional (3D) laser scanning and geographical information system (GIS) techniques for thematic mapping of facade stonework of St. Nicholas Church (Pisa, Italy). Int J Archit Herit 10(1):9–19 38. Lachat E, Landes T, Grussenmeyer P (2017) Performance investigation of a handheld 3D scanner to define good practices for small artefact 3D modeling. Int Arch Photogramm Remote Sens Spat Inf Sci XLII-2/W5:427–434. https://doi.org/10.5194/isprs-archives-XLII-2-W5-4272017 39. Applbaum N, Applbaum YH (2005) The use of medical computed tomography (CT) imaging in the study of ceramic and clay archaeological artifacts from the ancient near east. In: X-rays for archaeology. Springer, pp 231–245 40. Re A, Giudice AL, Nervo M, Buscaglia P, Luciani P, Borla M, Greco C (2016) The importance of tomography studying wooden artefacts: a comparison with radiography in the case of a coffin lid from ancient Egypt. Int J Con Sci 7(2):935–944 41. Kobayashi K, Hwang S-W, Okochi T, Lee W-H, Sugiyama J (2019) Non-destructive method for wood identification using conventional x-ray computed tomography data. J Cult Herit 38:88–93 42. Wilson P, Williams MA, Warnett JM, Attridge A, Ketchum H, Hay J, Smith MP (2017) Utilizing x-ray computed tomography for heritage conservation: the case of Megalosaurus bucklandii. In: 2017 IEEE international instrumentation and measurement technology conference (I2MTC). IEEE, pp 1–5 43. Murphy C, Fuller DQ, Stevens C, Gregory T, Parracho Silva F, Dal Martello R, Song J, Bodey AJ, Rau C (2019) Looking beyond the surface: use of high resolution x-ray computed tomography on archaeobotanical remains. Interdiscip Archaeol 10(1):7–18. https://doi.org/10.24916/ iansa.2019.1.1 44. Kirsch S (2019) Computed tomography as a tool for archiving ethnomusicological objects. In: Computational phonogram archiving. Springer, pp 305–319 45. Herman GT (2009) Fundamentals of computerized tomography: image reconstruction from projection. Springer 46. Edholm P, Herman G (1987) Linograms in image reconstruction from projections. Trans Med Imaging 6(4):301–307. https://doi.org/10.1109/TMI.1987.4307847 47. Hounsfield GN (1973) Computerized transverse axial scanning (tomography): part 1. Description of system. Br J Radiol 46(552):1016–1022 48. Vergnieux R (2011) Archaeological research and 3d models (restitution, validation and simulation). Virtual Archaeol Rev 2(4):39–43 49. Madrazo L (1994) Durand and the science of architecture. J Archit Educ 48(1):12–24 50. Palermo PC, Ponzini D (2010) Spatial planning and urban development: critical perspectives, vol 10. Springer Science & Business Media

48

eXtended Reality for Cultural Heritage

1435

51. Gonizzi Barsanti S, Remondino F, Visintini D (2012) Photogrammetry and laser scanning for archaeological site 3D modeling–some critical issues. In: Roberto V, Fozzati L (eds) Proceedings of the 2nd workshop on ‘the new technologies for Aquileia’. http://ceur-ws.org/Vol948/ 52. Stewart Z (2018) Tending the architectural corpus: a prehistory. J Archit Educ 72(2):290–293 53. Bressani, M. (2014). Architecture and the Historical Imagination: Eugène-Emmanuel Violletle-Duc, 1814–1879 (1st ed.). Routledge. https://doi.org/10.4324/9781315567679 54. Fuchs A (2006) Outils numériques pour le relevé architectural et la restitution archéologique. Theses, Université Henri Poincaré – Nancy I 55. Bennoui-Ladraa B, Chennaoui Y (2018) Use of photogrammetry for digital surveying, documentation and communication of the Cultural Heritage. example regarding virtual reconstruction of the access doors for the nameless temple of tipasa (algeria). Stud Digital Herit 2(2):121–137 56. Dawn S, Biswas P (2018) Technologies and methods for 3d reconstruction in archaeology. In: International symposium on signal processing and intelligent recognition systems. Springer, pp 443–453 57. Santacana i Mestre J, Carme Belarte Franco M (2005) Idées et débats de la restitution en archéologie, problèmes généraux concernant la restitution en archéologie. Centre des Monuments Nationaux, Editions du Patrimoine, pp 6–10 58. Gosselin LP (2014) Analyse des processus de recherche et de modélisation du patrimoine à partir de deux cas: l’écluse saint-gabriel du canal de lachine et la gare montreal and southern counties railways co. Masters’ thesis, Université de Sherbrooke 59. Huggett J (2015) A manifesto for an introspective digital archaeology. Open Archaeol 1(1): 86–95 60. Bekele MK, Pierdicca R, Frontoni E, Malinverni ES, Gain J (2018) A survey of augmented, virtual, and mixed reality for Cultural Heritage. J Comput Cult Herit 11(2):1–36. https://doi.org/ 10.1145/3145534 61. Carrozzino M, Voinea G-D, Duguleană M, Boboc RG, Bergamasco M (2019) Comparing innovative XR systems in Cultural Heritage. A case study. Int Arch Photogramm Remote Sens Spat Inf Sci XLII-2/W11:373–378. https://doi.org/10.5194/isprs-archives-XLII-2-W11373-2019 62. Ch’ng E, Li Y, Cai S, Leow F-T (2020) The effects of vr environments on the acceptance, experience, and expectations of Cultural Heritage learning. J Comput Cult Herit (JOCCH) 13(1):1–21 63. Cruz-Neira C, Sandin DJ, DeFanti TA, Kenyon RV, Hart JC (1992) The cave: audio visual experience automatic virtual environment. Commun ACM 35(6):64–72 64. Gaugne R, Gouranton V, Dumont G, Chauffaut A, Arnaldi B (2014) Immersia, an open immersive infrastructure: doing archaeology in virtual reality. Archeologia e Calcolatori, supplemento 5:1–10 65. Katsouri I, Tzanavari A, Herakleous K, Poullis C (2015) Visualizing and assessing hypotheses for marine archaeology in a VR CAVE environment. J Comput Cult Herit 8(2):1–18. https://doi. org/10.1145/2665072 66. Skarlatos D, Agrafiotis P, Balogh T, Bruno F, Castro F, Petriaggi BD, Demesticha S, Doulamis A, Drap P, Georgopoulos A, Kikillos F, Kyriakidis P, Liarokapis F, Poullis C, Rizvic S (2016) Project imareculture: advanced VR, immersive serious games and augmented reality as tools to raise awareness and access to European underwater Cultural Heritage. In: Ioannides M, Fink E, Moropoulou A, Hagedorn-Saupe M, Fresa A, Liestøl G, Rajcic V, Grussenmeyer P (eds) Digital heritage. Progress in Cultural Heritage: documentation, preservation, and protection. Springer International Publishing, Cham, pp 805–813 67. Lécuyer F, Gouranton V, Gaugne R, Nicolas T, Marchand G, Arnaldi B INSIDE interactive and non-destructive solution for introspection in digital environments. In: Digital heritage 2018 – 3rd international congress & expo, IEEE, new realities: authenticity & automation in the digital age, San Francisco, United States, October 2018, pp 1–4

1436

R. Gaugne et al.

68. Gaugne R, Petit Q, Nicolas T, Otsuki M, Gouranton V (2019) Evaluation of a mixed reality based method for archaeological excavation support. In: ICAT-EGVE: international conference on artificial reality and telexistence & Eurographics symposium on virtual environments, 2019, pp 1–8 69. Bimber O, Raskar R (2005) Spatial augmented reality: merging real and virtual worlds. A K Peters/CRC Press 70. Kurth P, Lange V, Siegl C, Stamminger M, Bauer F (2018) Auto-calibration for dynamic multiprojection mapping on arbitrary surfaces. IEEE Trans Vis Comput Graph 24(11):2886–2894. https://doi.org/10.1109/TVCG.2018.2868530 71. Ridel B, Reuter P, Laviole J, Mellado N, Couture N, Granier X (2014) The revealing flashlight: interactive spatial augmented reality for detail exploration of Cultural Heritage artifacts. J Comput Cult Herit 7(2):6:1–6:18 72. Gaugne R, Porcier S, Nicolas T, Coulon F, Hays O, Gouranton V A digital introspection of a mummy cat. In: Digital heritage, 3rd international congress & expo, new realities: authenticity & automation in the digital age, IEEE, San Francisco, 2018, pp 1–8 73. Bowman DA, Johnson DB, Hodges LF (1999) Testbed evaluation of virtual environment interaction techniques. In: Proceedings of the ACM symposium on virtual reality software and technology, VRST‘99. ACM, New York, pp 26–33 74. Hain V, Hajtmanek R (2019) Industrial heritage education and user tracking in virtual reality. In: Virtual reality. IntechOpen 75. Barreau J-B, Nouviale F, Gaugne R, Bernard Y, Llinares S, Gouranton V (2015) An immersive virtual sailing on the 18th-century ship Le Boullongne. Presence Teleop Virt Environ 24(3):201–219. https://doi.org/10.1162/PRES_a_00231 76. Chheang V, Saalfeld P, Huber T, Huettl F, Kneist W, Preim B, Hansen C (2019) Collaborative virtual reality for laparoscopic liver surgery training. In: IEEE international conference on artificial intelligence and virtual reality (AIVR) 77. Monsky WL, James R, Seslar SS (2019) Virtual and augmented reality applications in medicine and surgery-the fantastic voyage is here. Anat Physiol 9(1):313 78. Claude G, Gouranton V, Caillaud B, Gibaud B, Arnaldi B, Jannin P (2016) Synthesis and simulation of surgical process models. Stud Health Technol Inform 220:63–70 79. Roldán JJ, Crespo E, Martín-Barrio A, Peña-Tapia E, Barrientos A (2019) A training system for industry 4.0 operators in complex assemblies based on virtual reality and process mining. Robot Comput Integr Manuf 59:305–316 80. Santamaría-Bonfil G, Ibáñez MB, Pérez-Ramírez M, Arroyo-Figueroa G, Martínez-Álvarez F (2020) Learning analytics for student modeling in virtual reality training systems: lineworkers case. Comput Educ 151. https://doi.org/10.1016/j.compedu.2020.103871 81. R. B. Berthelot, T. Lopez, F. Nouviale, V. Gouranton, B. Arnaldi. Corvette: collaborative environment for technical training and experiment. In: 2014 IEEE virtual reality (VR), March 2014, pp 145–145 82. Cousseau F, Bernard Y, Petit Q, Barreau J-B, Gaugne R, Quesnel L Relever les élévations d’une architecture mégalithique: développements méthodologiques. In: GMPCA 2017 – XXIème Colloque international du Groupe des Méthodes Pluridisciplinaires Contribuant à l’Archéologie, Rennes, France, April 2017. CNRS UMR6566 CReAAH 83. Barreau J-B, Gaugne R, Bernard Y, Le Cloirec G, Gouranton V (2013) The west digital conservatory of archaelogical heritage project. In: DH, France, 2013, pp 1–8 84. Barreau J-B, Petit Q, Bernard Y, Auger R, Le Roux Y, Gaugne R, Gouranton V 3D reconstitution of the Loyola sugar plantation and virtual reality applications. In: Proceedings of computer applications & quantitative methods in archaeology, CAA, Sienna, Italy, 2015 85. Kreylos O, Bawden GW, Kellogg LH (2008) Immersive visualization and analysis of lidar data. In: International symposium on visual computing. Springer, pp 846–855 86. Barreau J-B, Gaugne R, Gouranton V (2017) Immersive point cloud manipulation for Cultural Heritage documentation. ERCIM News:33–34 87. Gaugne R, Petit Q, Barreau J-B, Gouranton V (2019) Interactive and immersive tools for point clouds in archaeology. In: ICAT-EGVE: international conference on artificial reality and telexistence & Eurographics symposium on virtual environments, 2019, pp 1–8

48

eXtended Reality for Cultural Heritage

1437

88. Nicolas T, Gaugne R, Tavernier C, Gouranton V, Arnaldi B (2014) Preservative approach to study encased archaeological artefacts. In: Ioannides M, Magnenat-Thalmann N, Fink E, Žarnić R, Yen A-Y, Quak E (eds) Digital heritage. progress in Cultural Heritage: documentation, preservation, and protection: 5th international conference, EuroMed 2014, Limassol, Cyprus, November 3–8, 2014. Proceedings. Springer International Publishing, Cham, pp 332–341 89. Nicolas T, Gaugne R, Tavernier C, Gouranton V, Arnaldi B (2016) Internal 3D printing of intricate structures. In: Ioannides M, Fink E, Moropoulou A, Hagedorn-Saupe M, Fresa A, Liestøl G, Rajcic V, Grussenmeyer P (eds) 6th international conference on culturage heritage – EuroMed 2016, volume 10058 of Lecture notes in computer science, Nicosia, Cyprus, October 2016, pp 432–441 90. Barreau J-B, Gaugne R, Olivier A-H, Llinares S, Gouranton V (2019) Reconstitution de la vie à bord d’un navire de la Compagnie des Indes Orientales au 18e siècle. In Situ: Revue des Patrimoines 39(2) 91. Nicolas T, Gaugne R, Tavernier C, Millet E, Bernadet R, Gouranton V (2018) Lift the veil of the block samples from the Warcq chariot burial with 3D digital technologies. In: Digital heritage 2018, 3rd international congress & expo, new realities: authenticity & automation in the digital age, San Francisco. IEEE, pp 1–8. https://doi.org/10.1109/DigitalHeritage.2018.8810036

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural Heritage María Mercedes Morita, Daniel Alejandro Loaiza Carvajal, and Gabriel Mario Bilmes

Contents 49.1 49.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Photogrammetry with Structure from Motion (SfM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49.2.1 Basics of SfM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49.2.2 Methods and Instrument Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49.2.3 SfM Software Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49.2.4 Potentials and Limitations of SFM Photogrammetry . . . . . . . . . . . . . . . . . . . 49.3 A 3D Digitization Solution for Low-Budget Public Museums: Mu3D System . . . 49.3.1 The Mu3D System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49.3.2 Outline for a 3D Digitization Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1440 1441 1441 1444 1448 1450 1451 1452 1455 1458 1459

Abstract

With the contribution of new technologies, the 3D digitization of museum collections opens great possibilities to perform conservation, research, education, and dissemination tasks in a more versatile, accessible, and attractive way. Although these applications have been carried out throughout the last decade in some institutions, the use of these technologies in public museums is still very limited. In this context, there is a demand for new photogrammetric systems for 3D digitization of cultural heritage. In this chapter, the state of the art of M. M. Morita (*) · D. A. Loaiza Carvajal Laboratorio de Ablación Láser, Fotofísica e Imágenes 3D, Centro de Investigaciones Ópticas (CONICET-CIC-UNLP), Gonnet, Provincia de Buenos Aires, Argentina e-mail: [email protected] G. M. Bilmes Laboratorio de Ablación Láser, Fotofísica e Imágenes 3D, Centro de Investigaciones Ópticas (CONICET-CIC-UNLP), Gonnet, Provincia de Buenos Aires, Argentina Departamento de Ciencias Básicas, Facultad de Ingeniería, Universidad Nacional de La Plata, La Plata, Provincia de Buenos Aires, Argentina e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_49

1439

1440

M. M. Morita et al.

photogrammetry with Structure from Motion as well as methods, instrumental setup, and software systems are reviewed. Based on this technique, an example of an easy low-cost 3D digitization system for low-budget public museums is presented. Their applications in the local public sphere are described, and a work plan for 3D digitization projects is proposed to be implemented in public museums that do not have easy access to this type of technologies.

49.1

Introduction

Museums keep and exhibit cultural heritage for conservation, dissemination, education, and research purposes. With the contribution of new technologies, the 3D digitization of the collections opens great possibilities to carry out these tasks in a more versatile, accessible, and attractive way. The exact virtual 3D reconstruction of a real object is known as Captured Reality (CR). The 3D image acquisition techniques most commonly used for CR in cultural heritage are laser or structured light (SL) scanning and digital photogrammetry [1–9]. Since recent years, the results obtained with digital photogrammetry by means of using the Structure from Motion (SfM) method [10] and appropriate image processing software are leaving behind the traditional scanning techniques, with the advantage of a lower cost, easy access, and better usability. This new option allows a really massive access to the use of the 3D resource, and it is beginning to impact very strongly in diverse areas, such as cultural heritage [11–13]. 3D models are increasingly used in the museums and research institutions worldwide for documentation of marks, or features in an object that are not detectable to a naked eye [14–16]; for monitoring the state of conservation of an artifact or site [17, 18]; for carrying out virtual restoration projects [19–21]; and for dissemination [22, 23]. They also allow real-scale measurements without a need of physical manipulation and the production of copies from originals [24]. Although these applications have been carried out for the last decade in some institutions, the use of these technologies in public museums is still very limited. In these cases, documentation and dissemination systems are generally photographic and do not include CR 3D models. One of the reasons could be the handling of the 3D imaging techniques, since the museum’s staff are generally not trained on how to use them. Another reason may be the high costs of the instrumentation commonly required for the 3D scanning techniques. In this context, there is a demand for new photogrammetric systems for 3D digitization of cultural heritage. Therefore, the Laboratory of Laser Ablation, Photophysics and 3D Imaging (LALFI) of the Centro de Investigaciones Ópticas (CIOp) in Argentina, has been studying the use of 3D technology in cultural heritage and developing easy low-cost 3D digitization systems for low-budget public museums. In this chapter, Mu3D, an example of such developments, will be described. The Mu3D digitization system consists of a photographic camera for image acquisition, a computer for image processing, SfM open-source software, and a friendly graphical user interface (GUI). This GUI was created by LALFI for museums’ staff in order to facilitate the interaction with

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1441

complex software involved in the 3D reconstruction process. LALFI is currently carrying out a project which aims to digitize in 3D the collections of low-budget regional museums. The Mu3D system is transferred to the institutions, by training their staff in the use of this technology and in order to offer assistance for the implementation of the corresponding pieces of software and conservation-related actions. In this chapter, the SfM photogrammetry technique is described. A summarizing guide on the methods, instrumental setup, and software solutions is provided to apply this technique to cultural heritage. In Sect. 49.3 of the chapter, Mu3D is laid out and proposed for low-budget public museums. The technical aspects and basic operations of the Mu3D system are described, as well as a work plan for a 3D digitization project, which can be replicated in public museums that do not have easy access to these technologies.

49.2

Photogrammetry with Structure from Motion (SfM)

49.2.1 Basics of SfM Photogrammetry is a technique based on the stereoscopy and the parallax principle that allows to measure lengths in an object, terrain or scene, and to obtain 3D information of the geometry by using two or more photographs with different lines of view of the same object or scene. Traditionally, the procedure required metric cameras (whose optical characteristics are well known), and also know the location of ground control points (regarding aerial photogrammetry) [25, 26]. With the contribution of computer vision algorithms, photogrammetry has evolved in an automatic and faster technique with the capacity of processing a huge amount of data [27], which is called Structure from Motion (SfM) [28]. There are plenty of pieces of software to perform 3D reconstruction. They have in common a sequential processing pipeline with an iterative reconstruction component, which use specific algorithms. It starts with the feature extraction and matching of a set of images of the object, followed by a geometric verification, scene points triangulation, and a reconstruction refinement using Bundle Adjustment [29, 30]. Currently, SfM photogrammetry is one of the most suitable 3D imaging techniques for the digitization of objects and scenes in different fields of study: geology, criminalistics, medicine, archaeology, and cultural heritage, among others. It has the advantage, over traditional photogrammetry, that any standard camera can be used and that the alignment can be performed on randomly positioned images. In comparison to other 3D acquisition techniques, such as laser or SL scanning, it does not require a complex setup, and it can be easily implemented by museum’s staff who have not been previously trained in image processing techniques. In addition, digital photogrammetry can reach the same resolution as the laser or SL scanning using low-cost equipment [6, 31, 32]. An SfM piece of software takes as input a set of images and generates a point cloud (3D coordinates) of the scene, the position from which each photo has been

1442

M. M. Morita et al.

taken and the optical parameters of the camera. In practice, one should take photos from different positions around the object, ensuring that each feature of it is visible in at least three photographs. When only two images are available, the algorithms may have errors that make 3D reconstruction impossible. To obtain better accuracy, more images should be used [33, 34]. In order to achieve the highest level of detail in the final 3D model, the images should have high resolution, so that more features can be detected and differentiated. A feature detection algorithm, e.g., Scale-invariant feature transform (SIFT), is used to detect the same feature between different views of the same object or scene, regardless of its scale, rotation, and illumination [35, 36] (Fig. 49.1). The algorithm will detect hundreds of points in each image, and compare them with those of the other images, to find matching pairs. If enough points are found, the computer will calculate the intrinsic parameters of the used camera. This information is essential in order to make up for any possible distortion of the lens and in order to determine the position of the focal point of each image. Afterward, the images are aligned with respect to each other through a process called “feature-based alignment” [37]. This is based on the principle of the intersection of the rays; using at least three images with a joint point in common, a ray is projected from the focal point of each image, through the detected feature points. The place where these rays intersect, then, determines the 3D coordinate of the feature point detected (Fig. 49.2). When this process is repeated for all points in the data set, the result is a sparse point cloud, which is an approximation of the scene’s 3D model. Figure 49.3 shows a sparse point cloud and the camera positions of the object of Fig. 49.1. This partial result can be processed to create a denser point cloud by finding additional junction points between the images. Since in this instance the camera’s original positions and optical characteristics are known, the computer can determine the overlapping of each image and, thus, reduce the area for searching common features again. Additional features detected in at least three images can be inserted back into the 3D point cloud using the principle of ray intersection. However, in this step, even if a point is detected in only two images, its 3D coordinates can be determined by means of using the parallax principle, as it happens in traditional

Fig. 49.1 Two images of the same object taken in different positions. The key points detected by SIFT (green dots) and the correspondences between those features (red lines) are shown. The object is a head sculpture from the Museo Provincial de Bellas Artes Emilio Pettoruti, La Plata, Argentina

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1443

Fig. 49.2 Projection and intersection of the rays corresponding to each key point to determine the 3D coordinate of that feature

Fig. 49.3 Representation of the camera positions computed during the 3D reconstruction of a head sculpture. The sparse point cloud can be seen in the center

photogrammetry. If all these additional points are added to the existing 3D model, the result is a much denser cloud of points (Fig. 49.4a) [27]. There are many algorithms available to generate dense point clouds [38], so different software systems may use different implementations.

1444

M. M. Morita et al.

Fig. 49.4 Dense point cloud (a). Mesh (b). Textured mesh (c)

By an interpolation between the points of the dense cloud, a 3D mesh can be obtained (Fig. 49.4b). Finally, the original images can then be projected on that mesh in a texture mapping process to create the final texture of the 3D model (Fig. 49.4c). Another way is by transferring the color of the points to the mesh. Surface and texture reconstruction filters [39, 40] and 3D processing software systems, such as OpenMVS [41] and MeshLab [42], can be used for these purposes [43–45]. The texturing technique chosen will depend on the purpose of the use of the 3D model. The accuracy of a 3D recording can vary significantly, since it depends on several interrelated factors: the number, resolution and quality of the images used, the size of the object, the degree of overlapping, the mismatches of the software processing, and the geometric arrangement of the images in relation to the object and between them.

49.2.2 Methods and Instrument Setup The following general recommendations can be taken to obtain suitable images for SfM photogrammetry: • A wide range of positions and angles, with an overlap greater than 70%. • A fixed focal length and a depth of field that allows most of the object to be in focus. • The illumination should be diffuse and even, making sure that the images are very clear in order to obtain more detectable features. • High resolution of the images. Nevertheless, this is not always related to the size of the image. Big sizes can significantly increase processing time. It is better to have a small sharp image rather than a big blurred image. • The object should occupy as much of the space as possible, so that the background or other elements do not compete with its features.

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1445

The use of a camera with manual mode is convenient, as it enables the user to control the exposure, focus, shutter speed, aperture, and their effect on the final image. As the diaphragm aperture influences in the depth of field, in order to perform SfM photogrammetry, it is better to have good illumination in the scene rather than using such large apertures. Apertures between f/8 and f/16 (quite closed) are recommended, as well as a low sensibility to light. However, this implies the need for an extended time exposure. Thus, the use of a tripod and remote shooting is very appropriate. Although lower focal lengths mean greater depth of field, it is recommended to reduce the aperture and use bigger focal lengths, in order to avoid optical and perspective distortions. It is suggested the use focal lengths of 40–55 mm, as well as distancing from the object approximately 1 or 2 m (for objects up to 1 m). Also, it is important to control the exposure, in such a way that the features of the selected object area are visible and detectable in each shot. The spot metering of the camera may be very useful. Finally, it is recommended that all images are inspected before being input into the SfM workflow. Poor-quality images, out of focus or blurred, should be eliminated. The problematic areas of the images can, in some software systems, be masked to remove features that do not need to be reconstructed, e.g., a background with many detectable features that can compete with the object. This reduces the time of the feature detection process. In regard to the overlapping of the input images, oblique and convergent images can be used as well as parallel images. Convergent photographs are low oblique photos in which camera axis converge toward one another. Using a convergent image configuration may be more convenient, as it reduces the systematic errors caused by inaccurate estimation of lens distortion characteristics [46, 47], and may also provide more overlapping of image pairs (almost 100%). In aerial photogrammetry, the use of drones and current photogrammetric software allows the use of oblique images, rather than flying in the more typical overlapping swaths used for aerial mapping. The addition of oblique images to a typical vertical set can significantly increase the accuracy of the results [47–49]. In addition, a greater redundancy of images may result in a higher-quality model. Re-photographing the same area by rotating the camera 90
and 180
can also be useful if the processing time is not significant. It is worth mentioning that, in aerial or large-scale photogrammetry, one of the parameters for determining the resolution of the input pictures of SfM photogrammetry is the Ground Sampling Distance (GSD). GSD is the distance between two consecutive pixel centers measured on the ground [50]. The GSD is related to the flight height: the higher the altitude of the flight, the bigger the GSD value. Another strategy used in this type of photogrammetry is to use control points data. If it is available, the whole process of image alignment can be optimized. On the other hand, one of the most important factors to consider is the layout of the object in relation to the camera shots and the environment. In general, two methods can be used to take the photos:

1446

M. M. Morita et al.

Fig. 49.5 Representation of the position of the camera in relation to the object. The camera should be placed at least in three different heights (which will determine three rings around the object) and three different orientation angles

1. The user moves around the static object. 2. The object rotates on its axis in front of a fixed camera. In this case, a plane background is necessary. The method used will depend on the size and portability of the object, as well as lighting conditions, space limitations, and the scene background. Either way, camera positions should be as in Fig. 49.5, i.e., the camera should be placed, at least, in three different heights (which will determine three rings around the object) and three different orientation angles; and each feature of the object should be present in at least three or more photos in each ring. In method 1, the normal condition is that the background changes from image to image in every shot, logically, by the displacement of the camera; this means that some elements of the background, such as walls, furniture, or plants, will also have some overlap. In this case, the resultant sparse point cloud will contain points of the background. This can be seen as accumulations with different densities depending on the number of features detected, but irrelevant for the 3D reconstruction of the main object. It is important that the object remains still during all the shots and to avoid high contrast shadows. If illumination can be controlled, method 2 is the best option. For this method, the simplest setup is: placing the object on a surface that allows it to rotate; using a tripod to keep the camera still, putting a plane background (without recognizable features), and two lamps left and right of the object just outside the field of view (Fig. 49.6).

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1447

Fig. 49.6 Setup for method 2. The object rotates; two light sources are placed for illuminating the object in a homogeneous way; and a plane background is placed behind the object, covering the table

This method is suitable for objects that can be moved to the photography area, and where lighting is artificial and can be controlled. It is the fastest and most systematic method once the scene is well arranged. As regards illumination, LED video lamps are an economical option and provide a lot of versatility. These usually have intensity controller, diffuser filters, option batteries or power supply, and ball joints to place them on any inclination, and can be fixed to a tripod or the same camera. Once they have been positioned and determined their angle, the intensity of the light can be adjusted until the object is evenly illuminated. Another way to obtain a uniform illumination on the object is with the aid of an annular flash, which is very suitable for small objects. The object should be raised to a certain height that allows the camera to take the photos also from below. If a base is used, then it must have the same diameter as the object’s own basis. It is also useful to have a turntable that simplifies the spin of the object by controlling the angle of rotation. To obtain exact measurements from the virtual model, it is advisable to set, near the real object, any other object whose ends are easily recognizable in the final 3D model and its exact real length is known.

1448

M. M. Morita et al.

Finally, in general, an object can almost never be digitized completely at once, as it is always supported by something (the basis, the floor, or a table). To obtain a 3D virtual reconstruction of an entire object, it is usually needed to make two independent reconstructions, e.g., the base and the upper part. These two models should then be joined in a 3D processing software, such as MeshLab. Placing small markers on the surface of the object can help in aligning the two parts correctly.

49.2.3 SfM Software Systems There is a wide variety of commercial, free, and open-source software packages for SfM. Most uses the basic algorithms of the technique (feature detection, Bundle Adjustment, and Multi-View Stereo). Open source solutions allow users to intervene in some specific instances of the 3D reconstruction process. Other software systems, aimed at the general public, either commercial (the most complete) or free systems, may be considered quite difficult because of the complexity of the process. Also, there are online applications with user account creation, yet these systems do not allow the user to intervene in the 3D reconstruction process. Maybe, the most popular all-to-one software solution for 3D reconstructions of monuments is Agisoft Metashape (formerly, Agisoft PhotoScan) [51–56]. Metashape is a commercial package with a semi-automated functionality of the software which turns out to be an efficient solution for users lacking technical knowledge. Metashape can also produce textured mapped meshes by blending some parts of the images, so that a photorealistic result can be achieved with low complexity 3D meshes [57]. The standard version offers photogrammetric triangulation, dense point cloud generation, 3D model generation, and texturing and panorama stitching. Conversely, the professional version offers working with ground control points for high accuracy surveying, measuring distances, areas, volumes, digital elevation model DSM/DTM export, georeferenced orthomosaic export, and multispectral imagery processing. The basic system requirements for working with Metashape are Quad-core Intel Core i7 CPU, Socket LGA 1150 or 1155 (Kaby Lake, Skylake, Broadwell, Haswell, Ivy Bridge, or Sandy Bridge); Motherboard LGA 1150 or 1155 model with 4 DDR3 slots and at least 1 PCI Express x16 slot; RAM DDR3–1600, 4  4 GB (16 GB total) or 4  8 GB (32 GB total); and GPU Nvidia GeForce GTX 980 or GeForce GTX 1080 (optional). This configuration should not present any complications if what is being processed is not extremely large data. Pix4D [58–61] is another commercial software that uses photogrammetry and computer vision algorithms to transform DSLR, fisheye, RGB, thermal, and multispectral images into 3D maps and 3D modeling. It operates on desktop, cloud, and mobile platforms, and the output products can be full color point clouds, orthomosaics, digital surface models (DSMs), digital elevation models (DEMs), 3d textured meshes (.ply, .fbx, .dxf, .obj, .pdf), index map, thermal maps, facade digital surface models, and facade orthomosaic, among others. Pix4D requires Windows 7, 8, 10, Server 2008, Server 2012, 64 bits (PC or Mac computers using Boot Camp); any CPU (Intel i5/ i7/ Xeon recommended); any GPU that is compatible with

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1449

OpenGL 3.2.; small projects (under 100 images at 14 MP): 4 GB RAM, 10 GB HDD Free Space.; medium projects (between 100 and 500 images at 14 MP): 8 GB RAM, 20 GB HDD Free Space; large projects (between 500 and 2000 images at 14 MP): 16 GB RAM, 40 GB HDD Free Space; very Large projects (over 2000 images at 14 MP): 16 GB RAM, 80 GB HDD Free Space. Since its release [62] in 2016, RealityCapture has become the software system selected for a wide range of industries, although most of its applications are in cultural heritage [63–66]. It is the commercial and improved version of CMPMVS free software system. It is 10 times faster than anything on the market and brings efficiency in the creation of virtual reality scenes, textured 3D meshes, orthographic projections, geo-referenced maps, and much more from images and/or laser scans completely automatically. RealityCapture runs on x64bit machines with at least 8GB of RAM, 64bit Windows 7/8/8.1/10, graphics card with Nvidia Cuda 2.0+ GPU and at least 1GB RAM. Previous studies have compared the previous software package from Agisoft and Pix4D in terrestrial systems and found minor variability in 3D reconstruction capability [67, 68], despite existing a vertical shift in height values of Agisoft digital surface model [69]. In Agisoft, the resulting mesh can be textured with minimum user effort by leaving the default settings [70]. In terms of affordability, fortunately, different licensing models are available for access to these software solutions. Renting these software packages for a couple of months can make them more affordable for those institutions with limited budgets, as long as the aim is to perform a specific short digitization work. Like open-source solutions, these commercial software systems allow local and controlled data processing. With regard to free solutions, one of the most popular throughout the last years has been VisualSFM [71], which is an application that runs fast by exploiting multicore parallelism for feature detection, feature matching, and Bundle Adjustment. For dense reconstruction, the PMVS/CMVS tool chain [72] can be used. The VisualSFM output can also be combined with many other tools, including CMPMVS [73], MVE [74], SURE [75] and MeshRecon [76]. VisualSFM can process a large number of images. Moreover, the user can intervene in the correspondence of features of the images, using his own sequence of pairs or using his own codes of feature detectors. The user can also adjust the speed and memory parameters used. In spite of the fact that its graphical interface is quite modest and it does not give the impression of executing complex processes, VisualSFM needs Gtk, SiftGPU, CUDA, GLEW, GLUT, DevIL, and PBA, among other packages, making its installation a non-trivial task. It also requires specific hardware for the different processes involved. Feature detection demands a high-quality GPU (ATI/Nvidia/ Intel). In particular, a large amount of GPU memory is needed. Conversely, Multicore Bundle Adjustment works on Nvidia CUDA or CPU. A very efficient pipeline used to be combining VisualSFM and CMPMVS in order to obtain high resolution colored and textured meshes. However, CMPMVS ceased to be developed in 2016, and it mutated into the commercial software Reality Capture. Another open source solution is COLMAP [77–79], which is a general-purpose Structure-from-Motion and Multi-View Stereo (MVS) pipeline with a graphical and

1450

M. M. Morita et al.

command-line interface. COLMAP goes through the SFM stages (feature detection and extraction; feature matching and geometric verification; sparse point cloud reconstruction) in different modules, which can be combined depending on the application. Then, MVS takes the output of SFM to compute depth and/or normal information for every pixel in an image. It fuses the depth and normal maps of multiple images in 3D and finally generates a dense point cloud of the scene. Surface reconstruction algorithms such as Poisson [39, 80] can then recover the 3D surface geometry of the scene, using the depth and normal information of the fused point cloud. COLMAP requires only few steps to do a standard reconstruction for a general user. For more experienced users, the program exposes many different parameters, only some of which are intuitive to a beginner. Once camera position and orientation have been computed either by OpenMVG [81] or COLMAP, Multiple View Stereo-vision algorithms could be used to compute a dense representation of the scene. OpenMVS [82] is one of the options to accomplish these tasks. The input is a set of camera poses plus the sparse pointcloud, and the output is a textured mesh. The main topics covered by this project are dense point-cloud reconstruction for obtaining a complete and accurate as possible point-cloud; mesh reconstruction for estimating a mesh surface that explains best the input point-cloud; mesh refinement for recovering all fine details; and mesh texturing for computing a sharp and accurate texture to color the mesh. Finally, Regard3D is one of the newest SfM free pieces of software with a simple workflow and a friendly GUI. Although it does not offer mesh or point cloud editing, it does offer surface reconstruction and texturing. Regard3D uses the AKAZE descriptor for feature detection [83] which is different from the others that usually use SIFT. With respect to feature-Detection-Description Time, although AKAZE is computationally efficient and more rotation invariant than SIFT, SIFT is consequently considered the most accurate algorithm [84]. A comparison study of different SFM software systems has been carried out in terms of efficiency, accuracy, constraints, and computation time [70]. Four 3D reconstruction free and open source software packages (VisualSFM v.0.5.26, Python Photogrammetry Toolbox with PPT GUI v.0.1, COLMAP v.3.1 and Regard3D v. 0.9.2) and the commercial software Agisoft’s Metashape v. 1.3.3 (considered as the ground truth) have been compared. It has been established that Regard3D is the best-rounded performer. COLMAP has also managed to capture the details without generating much noise, but it has taken three times longer for computation and it has not produced any texture. It is worth mentioning that none of these free open source applications which have been previously described offer any editing of point cloud or mesh, implying the use of an external application, such as MeshLab for further cleaning and editing 3D.

49.2.4 Potentials and Limitations of SFM Photogrammetry SfM photogrammetry has important advantages over traditional 3D scanning methods. It is cheaper, easier to use, and the whole process to produce a textured

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1451

3D mesh takes less time. The required instrumental is only a conventional camera and a computer. However, as it also happens with scanning techniques, if the quality of the meshes obtained is not good enough or adequate (incomplete surface, deformations, or variations in the color, often caused by uneven lighting), image processing techniques should be applied to improve the texture, clean or align them, involving a certain type of knowledge in image processing. Another advantage of SFM photogrammetry is its micro and macrophotogrammetry variant, using a microscope to record micrometric dimensions [85, 86]. Although it is known that laser and structured light scanning can be combined with microscopy, the configuration used to perform this task is very complex and difficult to handle. It is important to mention that SfM photogrammetry may produce 3D models with a minimal deformation with respect to the real object due to the distortion of the lens or software errors. A solution that offers greater accuracy for large-scale 3D registration is to use spatial reference systems, such as total stations or GPS systems, which allow calculating the coordinates in the field and transmitting them in digital form to the image processing software systems used later for the visualization of the measurements of the digitized object. However, these methods are only useful in large-scale 3D registration, such as geological formations or archaeological sites. The use of photogrammetry with Structure from Motion for the recording of small size objects has been poorly reported in the literature [87–90]. Most of the times, this type of object requires to be photographed at short distance, which is characterized by a low depth of field. There are solutions to this problem, as have been explained in Sect. 49.2.2., such as adjusting the aperture and the exposure time. Another option is to use focus stacking, which is an image processing technique that involves combining multiple images taken with different focus into a sharper image, where the entire object is focused. With respect to suitable objects for SfM photogrammetry, the following cases may result in a poor-quality 3D model, or even the cloud may not be generated: paintings, or very thin objects if it is intended to digitize both sides of it; kinetic pieces; bright or polished surfaces; totally dark objects, or with a monochrome plane surface; and hair, feathers, or small lattices. In these cases, complementary techniques, such as laser or SL scanning and pattern projection can be used to improve partially, or even, totally the results.

49.3

A 3D Digitization Solution for Low-Budget Public Museums: Mu3D System

Although 3D imaging technologies for cultural heritage are widely spread in the majority of public museums with budget limitations, the usual methods for recording and documenting are only the technical data record sheet and, at best, 2D photographs. One of the reasons why these institutions do not adopt such technologies may be that the most commonly used technique is laser scanning, which is, in most cases, quite expensive and difficult to access. On the other hand, in peripheral

1452

M. M. Morita et al.

countries there is a lack of articulation between the scientific-technological system and the community, especially in public institutions that preserve and exhibit cultural heritage. In these cases, public and institutional policies do not usually contemplate innovation, the appropriate use of new technologies, the acquisition of material resources, and the training of human resources. At best, museums’ staff is often aware of new technological advances, but they do not have funding, nor institutional support to implement them. In this context, the LALFI-CIOp has been studying the use of 3D technology in cultural heritage, and, since the beginning of 2017, it has been carrying out a project which has as its main objective the 3D digitization of the collections of the museums of the Province of Buenos Aires. The project, Digitization of cultural heritage by 3D images, is funded by the Comisión de Investigaciones Científicas de la Provincia de Buenos Aires and along with the Ministry of Culture of Argentina. In the framework of this project, a 3D digitization system, called Mu3D, has been developed. The project, and the Mu3D system itself, has a potential academic, educational, cultural, and social impact. From the academic and educational point of view, it is proposed to produce a technological innovation in the local disciplinary field of conservation by transferring the SfM photogrammetry technique for 3D digitalization to professionals in the museological field. From the cultural point of view, with the digitalization of the collections of the museums that have participated, it is expected to impact the cultural sector of the Province of Buenos Aires, especially in the relationship that the citizen establishes with the local heritage. From the social point of view, it is intended to facilitate the access to local cultural heritage by disseminating them digitally and expanding the number and diversity of the visiting public. In addition, with the new technologies that museums will be trained at, they are expected to acquire a new way of interacting with their collections and begin to rediscover new educational and cultural dissemination aspects. In order to transfer the Mu3D system to the public museums, the project contemplates the training of the museums’ staff and the support for the implementation of digitization strategies and conservation actions. Also, the project looks forward to the virtual access to the collections, e.g., allowing the community to visualize in 3D objects that are not exhibited. The Mu3D system and the project developed by LALFI are examples of the possibilities that SfM photogrammetry offers to digitize in 3D the collections of low-budget museums. In the following section, the Mu3D system is described, and a general outline of the project is proposed as a model for its replication.

49.3.1 The Mu3D System The Mu3D system consists of a computer and a camera and open source software for SfM and 3D image processing. In the last version of Mu3D, COLMAP is used to detect and match features, and to generate the sparse point cloud; OpenMVS, to obtain a denser point cloud and to generate a textured mesh; finally, Meshlab, to edit the 3D models, visualize them interactively and mapping deterioration and virtual

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1453

measurements, among other things. The SfM software tools (COLMAP and OpenMVS) are threaded in a graphical user interface (GUI) (Fig. 49.7) designed for the museums and institutions that participate in the aforementioned project in order to simplify the interaction with the complex pieces of software. We have found that COLMAP offers the best point detection, even though in some cases it may be a little slower than other software systems like VisualSFM. For that reason, COLMAP has been chosen to carry out the tasks which include point extraction and matching, and to create an NVM file with the sparse point cloud. Also, COLMAP offers the possibility to create a dense point cloud, but it has been found out that the OpenMVS algorithm has some advantages like time performance and more completed meshes. OpenMVS can also create the texture automatically. For that reason, it has been chosen as the software system in charge of densifying the sparse point cloud and creating the final mesh and texture. The Mu3D GUI has been developed in the Processing environment [91]. It is a set of screens that allow museums’ staff to easily create a 3D textured mesh. The development is focused on creating a friendly experience and integrates the entire process with just a few clicks (see outline in Fig. 49.8). It also works as a file manager for reconstruction projects. There are normally some minimum requirements to work with Mu3D software system: 8 GB RAM, Nvidia GTX 900, or 1000 series graphic card, and Intel® Core ™ i5 or better. Mu3D software system accepts any image format, but some complications have arisen when using images in PNG format, instead of JPG format, in OpenMVS. Photorealistic results have been achieved with the use of approximately 70 to 250 images, depending on the characteristics of the object. The recommended image resolution is between 3200 and 640 pixels for the larger side of the picture.

Fig. 49.7 Screenshot of the GUI of Mu3D (the current version is in Spanish)

Fig. 49.8 Mu3D software system workflow. Software processes (dashed line white), user actions (pink), alternatives (light blue), and results (yellow)

1454 M. M. Morita et al.

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1455

Table 49.1 Results of a 3D reconstruction performed with Mu3D system Result File name Colored sparse.nvm sparse point cloud

Colored sparse_dense.ply dense point cloud Mesh sparse_dense_mesh.ply Refined mesh Texturized mesh

sparse_dense_mesh_refine.ply sparse_dense_mesh_refine_texture.ply sparse_dense_mesh_refine_texture.png

Description Since this file contains the camera or rasters information, it can be used in Meshlab to create a texture for the 3D model, in the case that the texture obtained by Mu3D is not detailed enough Densification of the sparse point cloud

First mesh generated from the point could. Sometimes, it is a noisy 3D model It is an improved version of the first mesh, both in size and detail The refined mesh is used as an input to create a model that integrates the color of the real object

Fig. 49.9 Two 3D models visualized in a free online repository

The processing time depends on the hardware performance and suitability, as well as on the amount and pixels size of the pictures. A simple model with 70 images in a high-level computer can take 15 or 20 min; and with a higher number of pictures, it can take up to 3 or 4 h. After the entire process is finished, the user will have several files as the obtained results which are further described in Table 49.1. Figure 49.9 shows two 3D models visualized in the Sketchfab’s interface, and embedded in the project website.

49.3.2 Outline for a 3D Digitization Project The first instance is an agreement between a research and development institution (R&D), a government agency for cultural management, and the museums

1456

M. M. Morita et al.

(Fig. 49.10). In order to participate in the project, the museums should meet the following conditions: • Accept an institutional compromise to participate in the project. • Designate a suitable person for the digitization tasks who can also participate in the training course. • Select the collections and the estimated number of objects to be digitized in 1 year (≈10 objects). • Have a camera and a computer with an adequate processor, RAM, and graphics card for 3D image processing. In the next step, a training course is offered to the museums’ staff on the usage and capacities of the Mu3D system. In this course, theoretical contents and practical training in computers are given (16 h in total). The course also includes the tasks of the individual practical work which the trainees will have to carry out in order to pass the course. The contents developed in the course are: • • • •

Introduction to 3D recording for cultural heritage Basics of SFM photogrammetry and the Mu3D system Basics of photography The Mu3D software. Post-processing of the 3D model. Scale, alignment, and interventions. Exercises for further practice on PC • Discussion about the collections that each museum has proposed to digitize The R&D institution offers an open laboratory, so that the participants can train in SfM photogrammetry, test their own systems, and optimize the technique (photographic acquisition and data processing). The laboratory should be permanently available for participants who have completed the course, and it should bring the necessary support of the R&D institution. Although it should be something normal, public museums face many difficulties to access the purchase of mid-range computers. This is the case of the majority of the museums to which this proposal is addressed. They usually do not have the hardware requirements to work with 3D image processing systems. In any case, the possibility of attending a research laboratory, to use the computers, as in the project described above, allows overcoming this lack during the duration of a digitization project. 3D reconstruction usually demands some time, and the number of pieces that can be performed will depend on the duration of these processes and how the person in charge of the digitization works. Then, it is recommended to start digitizing “easy” objects, that is, with appropriate characteristics (see Sect. 49.2.4), in terms of materials and geometry, and which at the same time are interesting in terms of representativeness of the institution. In this sense, one of the most direct and easy dissemination policies is placing the 3D models into a website or social media. Whereas the contents of the course are the same for all museums, the versatility of the destinations of the applications of the 3D images allows each museum to respond

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

Fig. 49.10 Outline of a 3D digitization project proposed for low-budget museums

49

1457

1458

M. M. Morita et al.

to different institutional interests, in relation to the management of their collections: research, conservation, restoration, documentation, and dissemination. These interests should also be taken into account in the selection stage. In any case, the dissemination of collections may be the most accessible objective in a project that pursues also to introduce new technologies.

49.4

Conclusions

In this chapter, the possibilities that digital photogrammetry offers to 3D digitization of cultural heritage are presented. Easy and low-cost 3D digitization systems, based on photogrammetry with Structure from Motion, as the described Mu3D, can be developed and transferred to low-budget public museums. The hardware requirements of these systems include only a camera (it can be a cellphone camera) and a computer suitable for 3D image processing. These are elements that are assumed to be accessible to any public museum no matter their budget. Mu3D uses open source and commercial SfM software, which are becoming increasingly accessible to the cultural heritage professionals who are not specialized in handling 3D imaging. These software systems are also continuously improving the accuracy of the results, the texturing procedures, and adding tools for specific users. This type of technologies can be transferred to public museums by means of programs that articulate joint collaborations between museums, R&D institutions, and cultural management agencies. Based on the experience of the 3D digitalization project developed by LALFI in Argentina, an implementation proposal for low-budget museums and a general work plan have been presented. This proposal is not just a transfer of instruments or equipment, which often ends up being abandoned over time due to lack of technical knowledge and institutional commitment. Instead, what is proposed and offered to institutions is a permanent connection with the science and technology area. In addition, projects of this type encourage the use of open source software and low-cost instrumental solutions, especially suitable for public institutions. An aspect that should be considered when implementing projects of this type is that many museums do not have websites that allow them to place 3D models (and in many cases, they may not even have a website). Then, it is convenient to create a general repository in an institution that allows wide access to the digitized material. This repository can be centralized in a government institution and could be linked to the museum’s websites, available for anyone to access. In the example of the LALFI-CIOp project, the 3D models are first uploaded to the Sketchfab [92] public digital repository and then embedded in the project section of the CIOp institutional website [93]. Another aspect to consider is the possibility of using 3D reconstructions for the dissemination of cultural heritage. In this sense, immersive technologies, such as Virtual Reality [94] and Augmented Reality [95], are options that provide excellent alternatives. Nowadays, virtual museums offer either 2D images, or 3D models made with modeling techniques, instead of using Captured Reality. When VR and AR

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1459

technologies are combined with CR, it is possible to create virtual tours with elements from the real world. In this way, the public in general has access to environments which, otherwise, are inaccessible to them. Even though all of these technologies are relatively accessible to everybody, there are at the moment very few projects that aim to visualize art and cultural heritage by combining 3D modeling and CR [96], to generate virtual spaces that offer tours showing reconstructed objects in a realistic way. In this sense, the LALFI’s project implemented the virtual access to the museum collections, enabling the visualization and experimenting in 3D, e.g., objects that are not exhibited [97, 98]. Finally, augmented reality applications can also be developed, e.g., the augmented reality app, called MusAR, from LALFI that can be downloaded from the Google Play Store [99]. Acknowledgments This work has been supported by a PIT-AP-BA project of the Comisión de Investigaciones Científicas de la Provincia de Buenos Aires (CIC-BA) and by the Projects 11/I199 and I240 of Facultad de Ingeniería, Universidad Nacional de La Plata. M.M. Morita is Research Member at Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). D.A. Loaiza Carvajal is PhD fellow at Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). G.M. Bilmes is Research Member at Comisión de Investigaciones Científicas de la Provincia de Buenos Aires (CIC-BA) and Full Professor at Facultad de Ingeniería, Universidad Nacional de La Plata (FI-UNLP).

References 1. Remondino F, Rizzi A (2010) Reality-based 3D documentation of natural and cultural heritage sites – techniques, problems, and examples. Appl Geomat 2(3):85–100. https://doi.org/10.1007/ s12518-010-0025-x 2. Remondino F (2011) Heritage recording and 3D modeling with photogrammetry and 3D scanning. Remote Sens 3(6):1104–1138. https://doi.org/10.3390/rs3061104 3. Sansoni G, Trebeschi M, Docchio F (2009) State-of-the-art and applications of 3D imaging sensors in industry, cultural heritage, medicine, and criminal investigation. Sensors 9:568–601. https://doi.org/10.3390/s90100568 4. Schaich M (2009) 3D-Scanning-Technologien in der Bau- und Kunstdenkmalpflege und der archäologischen Feld- und Objektdokumentation. In: Faulstich E, Hahn-Weishaupt A (eds) Dokumentation und Innovation bei der Erfassung von Kulturgütern. Schriften des Bundesverbands freiberuflicher Kulturwissenschaftler 2. Online-publication: http://www.b-fk.de/webpub01/cnt/schaich.htm 5. Ioannides M, Fritsch D, Leissner J, Davies R, Remondino F, Caffo R (eds) (2012) Progress in cultural heritage preservation, 4th international conference, EuroMed 2012, Limassol, Cyprus, proceedings series: Lecture notes in computer science, vol 7616. Subseries: Information systems and applications, incl. Internet/Web, and HCI. 2012, XXV. Irschara 6. Morita M, Bilmes GM (2018) Applications of low-cost 3D imaging techniques for the documentation of heritage objects. Óptica Pura y Aplicada 51(2):50026:1–50026:11. https://doi.org/ 10.7149/OPA.51.2.50026 7. Yanagi H, Chikatsu H (2010) 3D modelling of small objects using macro lens in digital very close-range photogrammetry. ISPRS Arch 38:617–622 8. Remondino F, Spera MG, Nocerino E, Menna F, Nex F (2014) State of the art in high density image matching. Photogramm Rec 29(146):144–166 9. Stylianidis E, Remondino F (2016) 3D recording, documentation and management of cultural heritage. Whittles Publishing

1460

M. M. Morita et al.

10. Micheletti N, Chandler JH, Lane SN (2015) Structure from motion (SfM) photogrammetry. Photogramm Herit 2:1–12 11. Roncella R, Re C, Forlani G (2011) Performance evaluation of a structure and motion strategy in architecture and cultural heritage. Int Arch Photogramm Remote Sensing Spat Inf Sci 38:285–292 12. Chiabrando F, Donadio E, Rinaudo F (2015) SfM for orthophoto to generation: a winning approach for cultural heritage knowledge. Proc Int Arch Photogramm Remote Sensing Spat Inf Sci 40(5):91–98 13. Chiabrando F, Lingua A, Noardo F, Spanò A (2014) 3D modelling of trompe l’oeil decorated vaults using dense matching techniques. Ann Photogramm Remote Sensing Spat Inf Sci 2(5): 97–104 14. Fernández-Lozano J, Gutiérrez-Alonso G, Ruiz-Tejada MA, Criado-Valdés M (2017) 3D digital documentation and image enhancement integration into schematic rock art analysis and preservation: the Castrocontrigo Neolithic rock art (NW Spain). J Cult Herit 26:160–166. https:// doi.org/10.1016/j.culher.2017.01.008 15. Monna F, Esin Y, Magail J, Granjon L, Navarro N, Wilczek J, Saligny L, Couette S, Dumontet A, Chateau C (2018) Documenting carved stones by 3D modelling – example of Mongolian deer stones. J Cult Herit 34:116–128. https://doi.org/10.1016/j.culher.2018.04.021 16. Deger F, Mansouri A, Curdy P, Pedersen M, Hardeberg JY, Voisin Y (2016) Statistical analysis of engraving traces on a 3D digital model of prehistoric stone stelae. J Cult Herit 17:151–158. https://doi.org/10.1016/j.culher.2015.07.007 17. Lercari N (2019) Monitoring earthen archaeological heritage using multi-temporal terrestrial laser scanning and surface change detection. J Cult Herit 39:152–165. https://doi.org/10.1016/j. culher.2019.04.005 18. Palma G, Pingi P, Siotto E, Bellucci R, Guidi G, Scopigno R (2019) Deformation analysis of Leonardo da Vinci’s “Adorazione dei Magi” through temporal unrelated 3D digitization. J Cult Herit 38:174–185. https://doi.org/10.1016/j.culher.2018.11.001 19. Lanitis A, Stylianou G, Voutounos C (2012) Virtual restoration of faces appearing in byzantine icons. J Cult Herit 13(4):404–412. https://doi.org/10.1016/j.culher.2012.01.001 20. Fowles PS, Larson JH, Dean C, Solajic M (2003) The laser recording and virtual restoration of a wooden sculpture of Buddha. J Cult Herit 4(1):367–371. https://doi.org/10.1016/S12962074(02)01141-X 21. Arbace L, Sonnino E, Callieri M, Dellepiane M, Fabbri M, Iaccarino Idelson A, Scopigno R (2013) Innovative uses of 3D digital technologies to assist the restoration of a fragmented terracotta statue. J Cult Herit 14(4):332–345. https://doi.org/10.1016/j.culher.2012.06.008 22. Jiménez Fernández-Palacios B, Morabito D, Remondino F (2017) Access to complex realitybased 3D models using virtual reality solutions. J Cult Herit 23:40–48. https://doi.org/10.1016/ j.culher.2016.09.003 23. Magnani M, Guttorm A, Magnani N (2018) Three-dimensional, community-based heritage management of indigenous museum collections: archaeological ethnography, revitalization and repatriation at the Sámi Museum Siida. J Cult Herit 31:162–169. https://doi.org/10.1016/j. culher.2017.12.001 24. Balletti C, Ballarin M, Guerra F (2017) 3D printing: state of the art and future perspectives. J Cult Herit 26:172–182. https://doi.org/10.1016/j.culher.2017.02.010 25. Gómez Lahoz J (2010) Introducción a la fotogrametría digital: el método general de la fotogrametría digital. Departamento de Ingeniería Cartográfica y del Terreno, Escuela Politécnica Superior de Ávila 26. Moffitt FH, Mikhail EM (1980) Photogrammetry, 3rd edn. Harper & Row, New York 27. Van Damme T (2015) Computer vision photogrammetry for underwater archaeological site recording. PhD thesis, University of Southern Denmark 28. Snavely N, Seitz SN, Szeliski R (2008) Modeling the world from internet photo col- lections. Int J Comput Vis 80:189–210

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1461

29. Schönberger JL, Frahm J (2016) Structure-from-motion revisited. In: IEEE conference on computer vision and pattern recognition (CVPR), Las Vegas, IEEE, pp 4104–4113. https:// doi.org/10.1109/CVPR.2016.445 30. Triggs B, Mclauchlan PF, Hartley RI, Fitzgibbon AW (2000) Bundle adjustment – a modern synthesis. Vis Algorithms Theory Pract 34099:298–372. https://doi.org/10.1007/3-540-44480-7 31. Sapirstein P (2018) A high-precision photogrammetric recording system for small artifacts. J Cult Herit 31:33–45. https://doi.org/10.1016/j.culher.2017.10.011 32. Skarlatos D, Kiparissi S (2012) Comparison of laser scanning, photogrammetry and SFM-MVS pipeline applied in structures and artificial surfaces. SPRS Ann Photogramm Remote Sens Spatial Inf Sci I-3:299–304 33. Torres JC, de Haro J, Arroyo G, Romo C (2012) 3D digitization using structure from motion. CEIG – Spanish Comput Graph Conf 1–10 34. Corsini M, Dellepiane M, Ganovelli F, Gherardi R, Fusiello A, Scopigno R (2013) Fully automatic registration of image sets on approximate geometry. Int J Comput Vision 102(1–3):91–111. https:// doi.org/10.1007/s11263-012-0552-5 35. Lowe DG (1999) Object recognition from local scale-invariant features. In: Proceedings of international conference on computer vision, pp 1150–1157 36. Rey-Otero I, Delbracio M (2014) Anatomy of the SIFT method. Image Process On Line 4:370– 396. https://doi.org/10.5201/ipol.2014.82 37. Szeliski R (2010) Computer vision: algorithms and applications. Springer Science & Business Media 38. Furukawa Y, Hernandez C (2015) Multi-view stereo: a tutorial. Found Trends Comput Graph Vis 9(1–2):1–148. https://doi.org/10.1561/0600000052 39. Kazhdan M, Bolitho M, Hoppe H (2006) Poisson surface reconstruction. In: Polthier K, Sheffer A (eds) Eurographics symposium on geometry processing, Eurographics Association 40. Dellepiane M, Callieri M, Corsini M, Cignoni P, Scopigno R (2010) Improved color acquisition and mapping on 3D models via flash-based photography. ACM J Comput Cult Herit 2:1–20 41. OpenMVG: https://openmvg.readthedocs.io/en/latest/software/MVS/OpenMVS/ 42. MeshLab: http://www.meshlab.net 43. Cignoni P, Callieri M, Corsini M, Dellepiane M, Ganovelli F, Ranzuglia G (2008) MeshLab: an open-source mesh processing tool. In: Proc. Eurographics Italian chapter conference, pp 129–136 44. Callieri M, Cignoni P, Corsini M, Scopigno R (2008) Masked photo blending: mapping dense photographic dataset on high-resolution 3D models. Comput Graph 32(4):464–473 45. Corsini M, Dellepiane M, Ponchio F, Scopigno R (2009) Image-to-geometry registration: a mutual information method exploiting illumination-related geometric properties. Computer Graphics Forum 28(7):1755–1764. https://doi.org/10.1111/j.1467-8659.2009.01552.x 46. Chandler JH (2010) Tips for the effective use of close-range digital photogrammetry for the Earth sciences. ISPRS Commission V, Working Group V/6. Available at: http://isprsv6.lboro.ac. uk/tips.html. Accessed 29 Sept 2019 47. Wackrow R (2008) Spatial measurement with consumer grade digital cameras. PhD thesis, Loughborough University. Available at: https://dspace.lboro.ac.uk/dspace-jspui/handle/2134/ 3827. Accessed 29 Sept 2019 48. Nocerino E, Menna F, Remondino F, Saleri R (2013) Accuracy and block deformation analysis in automatic UAV and terrestrial photogrammetry: lesson learnt, ISPRS annals II-5/W1 203–8 49. Bedford J (2017) Photogrammetric applications for cultural heritage. Guidance for good practice. Historic England, Swindon 50. Aerial Survey Base (2019) What is GSD. https://www.aerial-survey-base.com/gsd-calculator/ what-is-gsd/ 51. Agisoft: https://www.agisoft.com/ 52. Abdelaziz M, Elsayed M (2019) Underwater photogrammetry digital surface model (DSM) of the submerged site of the ancient lighthouse near Qaitbay Fort in Alexandria, Egypt, Proc. the International Archives of the Photogrammetry, Remote Sensing and Spatial Information

1462

M. M. Morita et al.

Sciences, XLII-2/W10, 2019. Underwater 3D recording and modelling “a tool for modern applications and CH recording”, 2–3 May 2019, Limassol, Cyprus 53. Barelli L, Paolini P, Forti G (2017) The XII century towers, a benchmark of the Rome countryside almost cancelled: the safeguard plan by low cost UAV and terrestrial DSM photogrammetry surveying and 3D web GIS applications, Proc. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol XLII-2/W8, 2017. 5th international workshop LowCost 3D – sensors, algorithms, applications, 28–29 November 2017, Hamburg, Germany 54. Shafqat Malik U, Guidi G (2018) Massive 3D digitization of sculptures: methodological approaches for improving efficiency. In: Florence Heri-Tech – the future of heritage science and technologies. IOP Conf. Ser.: Mater. Sci. Eng. 364, p 012015 55. Remondino F, Nocerino E, Toschi I, Menna F (2017) A critical review of automated photogrammetric processing of large datasets. Int Arch Photogramm Remote Sens Spat Inf Sci – ISPRS Arch, International Society for Photogrammetry and Remote Sensing: 591–599. https:// doi.org/10.5194/isprs-archives-XLII-2-W5-591-2017 56. Becker S, Spranger M, Heinke F, Grunert S, Labudde D (2018) A comprehensive framework for high resolution image-based 3D modeling and documentation of crime scenes and disaster sites. Int J Adv Syst 11:1–12 57. Koutsoudis A, Vidmar B, Ioannakis G, Arnaoutoglou F, Pavlidis G, Chamzas C (2014) Multiimage 3D reconstruction data evaluation. J Cult Herit 15(1):73–79. https://doi.org/10.1016/j. culher.2012.12.003 58. Pix4d: https://www.pix4d.com 59. Murtiyoso A, Frussenmeyer P, Koehl M, Freville T (2016) Acquisition and Processing experiences of close-range UAV images for the 3d modeling of Hertitage Buildings. In: Ioannides M et al (eds) EuroMed 2016, part 1, LNCS 10058. Springer International Publishing AG, pp 420– 431. https://doi.org/10.1007/978-3-319-48496-9_34 60. Gagliolo S, Ausonio E, Federici B, Ferrando I, Passoni D, Sguerso D (2018) 3D cultural heritage documentation: a comparison between different photogrammetric software and their products. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 42(2). ISPRS TC II mid-term symposium “towards photogrammetry 2020”, 4–7 June 2018, Riva del Garda, Italy 61. Landauer J Towards automating drone flights for archaeological site documentation. Conference: International Aerial Archaeology Conference AARG 2018, Venice, Italy, 2018. Available at: ResearchGate 62. Reality Capture: https://www.capturingreality.com 63. Zamani: Preserving Sri Lanka’s heritage: https://www.capturingreality.com/Article-ZamaniSriLanka-Project 64. Marie-Antoinette’s hamlet in VR: https://www.capturingreality.com/Article-Marie-Antoinettehamlet-in-VR 65. CyArk and RealityCapture: https://www.capturingreality.com/Article-CyArk-and-RealityCapture 66. Factum Foundation and RealityCapture: https://www.capturingreality.com/Article-Factum Foundation-and-RC 67. Niederheiser R, Mokroš M, Lange J, Petschko H, Prasicek G, Elberink SO (2016) Deriving 3D point clouds from terrestrial photographs-comparison of different sensors and software. Int Arch Photogramm Remote Sens Spat Inf Sci ISPRS Arch 41:685–692 68. Burns JHR, Delparte D (2017) Comparison of commercial structure-from-motion photogrammety software used for underwater three-dimensional modeling of coral reef environments. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 42(2)/W3. 3D virtual reconstruction and visualization of complex architectures, 1–3 March 2017, Nafplio, Greece 69. Alidoost F, Arefi H (2017) Comparison of UAS-based photogrammetry software for 3D point cloud generation: a survey over a historical site, ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 4(4)/W4. 4th international GeoAdvances workshop, 14–15 October 2017, Safranbolu, Karabuk, Turkey

49

New Photogrammetric Systems for Easy low-Cost 3D Digitization of Cultural. . .

1463

70. Rahaman H, Champion E (2019) To 3D or not 3D: choosing a photogrammetry workflow for cultural heritage groups. Heritage 2(3):1835–1851. https://doi.org/10.3390/heritage2030112 71. VisualSFM: http://ccwu.me/vsfm/ 72. Furukawa Y, Ponce J (2007) Accurate, dense, and robust multi-view stereopsis. In: Proc. IEEE conference on computer vision and pattern recognition (CVPR), 17–22 June, Minneapolis, USA, pp 1–8 73. Jancosek M, Pajdla T (2011) Multi-view reconstruction preserving weakly-supported surfaces. In: Proc. IEEE conference on computer vision and pattern recognition. https://doi.org/10.1109/ CVPR.2011.599569 74. Fuhrmann S, Langguth F, Goesele M MVE – a multi-view reconstruction environment. In: Proc. Eurographics workshop on graphics and cultural heritage, Darmstadt, Germany, 2014 75. Rothermel M, Wenzel K, Fritsch D, Haala N SURE: photogrammetric surface reconstruction from imagery. In: Proc. LC3D workshop, Berlin, December 2012 76. Kang Z, Medioni G (2015) 3D urban reconstruction from wide area aerial surveillance video. In: Applications and Computer Vision Workshops (WACVW), IEEE Winte 77. COLMAP: https://colmap.github.io 78. Schönberger JL (2018) Robust methods for accurate and efficient 3D modeling from unstructured imagery. ETH Zürich 79. Bianco S, Ciocca G, Marelli D (2018) Evaluating the performance of structure from motion pipelines. J Imaging 4(8):98. https://doi.org/10.3390/jimaging4080098 80. Kazhdan M, Hoppe H (2013) Screened poisson surface reconstruction. ACM Trans Graph 32(3): 29, 13 pages. https://doi.org/10.1145/2487228.2487237 81. OpenMVG: https://openmvg.readthedocs.io/en/latest/ 82. OpenMVS: https://openmvg.readthedocs.io/en/latest/software/MVS/OpenMVS/ 83. Alcantarilla PF, Nuevo J, Bartoli A (2013) Fast explicit diffusion for accelerated features in nonlinear scale spaces, British Machine Vision Conference. Available at: http://www.bmva.org/ bmvc/2013/Papers/paper0013/paper0013.pdf 84. Tareen SAK, Saleem Z A comparative analysis of SIFT, SURF, KAZE, AKAZE, ORB, and BRISK. In: 2018 international conference on computing, mathematics and engineering technologies (iCoMET), Sukkur, 2018, pp 1–10. https://doi.org/10.1109/ICOMET.2018.8346440 85. Tiano P, Tapete D, Matteini M, Ceccaroni F The micro-photogrammetry: a new diagnostic tool for onsite monitoring of monumental surfaces. In: Proc. international workshop SMW08 In situ monitoring of monumental surface, 27 Oct 2008 86. Yanagi H, Chikatsu H (2010) 3D modeling of small objects using macro lens in digital very close-range photogrammetry. ISPRS Arch 38:617–622 87. Alby E, Smigiel E, Assali P, Grussenmeyer P, Kauffmann-Smigiel I Low cost solutions for dense point clouds of small objects: photomodeler scanner vs. David laserscanner, 22nd CIPA symposium, Kyoto, October 11–15, 2009 88. Samaan M, Héno R, Pierrot-Deseilligny M (2013) Close-range photogrammetric tools for small 3D archeological objects. Int Arch Photogramm Remote Sens Spat Inf Sci 40:549–553 89. Galantucci LM, Pesce M, Lavecchia F (2016) A powerful scanning methodology for 3D measurements of small parts with complex surfaces and sub millimetre-sized features, based on close range photogrammetry. Precis Eng 43:211–219 90. Marroquim R, Sá A, Rodrigues K, Balbio V, Zamorano R (2017) Digitising ivory artefacts at the National History Museum in Brazil, Eurographics workshop on graphics and cultural heritage 91. Processing: https://processing.org/reference/environment/ 92. Sketchfab: https://sketchfab.com 93. Mu3D Project: https://www.proyectosciop.unlp.edu.ar/mu3d/wp/ 94. Mazuryk T, Gervautz M (1999) Virtual reality – history, applications, technology and future. Institute of Computer Graphics. Vienna University of Technology, Austria. https://www.cg. tuwien.ac.at/research/publications/1996/mazuryk-1996-VRH/TR-186-2-96-06Paper.pdf 95. Milgram P, Takemura H, Utsumi A, Kishino F (1995) Augmented reality: a class of displays on the reality-virtuality continuum. In: Telemanipulator and telepresence technologies, vol 2351. International Society for Optics and Photonics, pp 282–293

1464

M. M. Morita et al.

96. Muñoz Morcillo J, Faion F, Zea A, Hanebeck UD (2016) E-Installation: synesthetic documentation of media art via telepresence technologies. In: Dan MB, Crăciun C (eds) Space and time visualization. Springer Nature, pp 173–191 97. Krause. Vestigios disponibles. Web version: https://www.proyectosciop.unlp.edu.ar/mu3d/Krause %20Web/ 98. Virtual Collections. Web version: https://www.proyectosciop.unlp.edu.ar/mu3d/Krause% 20Web/https://www.proyectosciop.unlp.edu.ar/mu3d/Colecciones%20Virtuales%20Web/ 99. MusAR augmented reality app: https://play.google.com/store/apps/details?id¼com.ciop.musar& hl¼en

Role of Digital Technologies, Robotics, and Augmented Realities

50

Paolo Salonia

Contents 50.1 50.2

Knowledge-Based Process in the Conservation Project . . . . . . . . . . . . . . . . . . . . . . . . . . Some Previous Activities in Using Enabling Technologies for Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.3 Heritage/Knowledge/Technology. Utopias and Dystopias . . . . . . . . . . . . . . . . . . . . . . . 50.4 Technological World and Future . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.5 Cultural Heritage, Identity Values, Memory in the Twenty-First Century . . . . . . . 50.6 Conclusion or Where Will Our Memory Be? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1467 1468 1471 1482 1486 1490 1492

Abstract

Several examples testify how the evolution in applied technologies to cultural heritage has improved significantly in the last decades in terms of accuracy and reliability of measurement, restitution, and management of the acquired data. This evolution is not always accompanied by the simplification of procedures, lowering of costs, and, more importantly, awareness-raising of the identity value of the investigated asset. In this article, some case studies where different types of technologies have been used are presented together with a recent study of applied robotics in cultural heritage. The aim is to construct an exhaustive picture of the technologies currently in use within the complex conservation process. To follow, a series of urgent reflections are introduced in order to stimulate, contribute, and increase the

P. Salonia (*) Institute of Cultural Heritage Sciences of Italian National Research Council – ISPC CNR (formerly Institute for Technologies Applied to Cultural Heritage – ITABC CNR) - ICOMOS IT (Advisor and Executive Board Member) - International Committee of Architectural Photogrammetry - CIPA (Expert Member), Rome, Italy e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_50

1465

1466

P. Salonia

necessary interdisciplinary and intercultural debate on the issue of safeguarding cultural assets as a driving force for civil progress, paying particular attention to 3D digital survey. Above all, the paper expands the issue to a wider analysis of the general and progressive anthropological transformations produced on human society by the speed of technological innovation. Finally, the paper will attempt to verify how much and what role the identity values expressed by cultural heritage will still exercise on technologically globalized cities and societies of the twenty-first century. Keywords

Anthropological transformation · Artificial Intelligence · Augmented reality · Authenticity · Big data · Communication · Computer graphics · Computer vision · Cultural heritage · Digital content · Digital divide · Digital Heritage · Digitalization · Dissonant heritage · Education/training · Emotional performance · Enabling technologies · Enjoyment · Ethical and esthetical category · Globalization · Hard science · Human and technological · Humanism · Humanities · Identity values · Image analysis · Infosphere · Interdisciplinarity · Internet of Games I · Internet of Things · Knowledge · Memory · Mixed reality · Monitoring · Multi-image-matching systems · Natural and artificial · Own memory · Programmatic conservation · Reproducibility · Resilience · Right to Heritage · Robotic · Serious game · Simplification · Smart historic cities · Standardization · Structure for Motion · Sustainable development · Technological divide · Technology · Touristification · Unmanned Aerial Vehicles · Virtual reality Abbreviations

AI CHARISMA CIPA CNR DARIAH EPOCH EU HUL ICOMOS ICT IoG IoT JPI SDGs SEAV SfM

Artificial Intelligence Cultural Heritage Advanced Research Infrastructures SMA Scientific Committee on Heritage Documentation Italian National Research Council Digital Research Infrastructure for the Arts and Humanities European Research Network on Excellence in Processing Open Cultural Heritage European Union Historic Urban Landscape International Council on Monuments and Sites Information and Communication Technology Internet of Games Internet of Things Joint Programming Initiative on Cultural Heritage United Nations Sustainable Development Goals Spanish Society of Virtual Archaeology Structure for Motion

50

Role of Digital Technologies, Robotics, and Augmented Realities

UAV UNESCO

50.1

Unmanned Aerial Vehicle United Nations Educational, Organization

Scientific,

1467

and

Cultural

Knowledge-Based Process in the Conservation Project

The project of conservation derives from the recognition of the values that belong to the investigated artifact. The methods and the criteria of the intervention should be defined after a preliminary dimensional, material, phenomenological characterization, as well as a historical-critical analysis of the intervention itself. Only in a vast and eco-systemic meaning of cultural heritage, it is possible to fully catch the value of a systematic approach to the entire process, considered as a complex system of interactions among different constantly cyclic phases, knowledge, conservation, monitoring, valorization, and fruition. This system constantly needs reliable, scientific documentation where the acquisition of data for the realization of informative systems and explorable 3D environments represents a key moment to test innovative survey systems integrated with traditional techniques. The feature of transversality of this knowledge phase is at the base of the concept of programmatic conservation. This approach goes beyond the discourse of intervention as an interruption of the cycle of life of an artifact. Conversely, this approach wants to harmonize itself in a co-evolutionary way with the cycle of life, intervening mildly, continuously, and sustainably. It is a complex strategy which requires new scenarios, where a combination of planning, research, innovation, and interdisciplinarity – and through which the event (whether more or less traumatic) of restoration (or of the simple overtime maintenance) – is substituted by a continuous process of knowledge, prevention, control, and planning of interventions. Within this process, the practice of monitoring represents a paramount tool. For definition, monitoring represents an articulate procedure, since it depends on the intrinsic characteristics of the artifact, the different and multiple environmental situations, complex phenomenologies for investigation, and diverse types of intervention, materials, and products. In order to be effective, the monitoring must have some requisites of continuity (to align itself with the time of decay), objectivity (to guarantee the repeatability of the acquisitions and the comparison of the data), and affordability (to ensure the necessary continuity and the operational frequency). In order to guarantee the above objectivity, it will be compulsory to establish data measurement protocols. These protocols should cover procedures and methods of data collection, analysis and management, and the transfer of heterogeneous data – which to be understood and utilized (not only in their uniqueness but especially in their interactive ensemble) require the creation of specific environments for the integration of data. Today, the methodologies – which integrate the primary aspects of the disciplines related to the conservation and valorization enhancement of cultural property

1468

P. Salonia

heritage with those of the Information and Communication Technology (ICT) [1] – significantly support the operations necessary to provide thorough contents to the phase of knowledge. In particular, the evolution of technology for the survey of the geometry represents the innovative element that has allowed the introduction of the morphologic metrical data at high informative density as an essential support to the construction of three-dimensional databases. These archives of the geometric memory of architecture and cultural property heritage demonstrate, over time, their useful application to research by art and architecture historians, and also to analyses on decay, safeguarding, enhancement, and support for potential stages of planning and restoration, or for the creation of reproductions for museum purposes. For several decades, the traditional practices have increasingly become more focused on the instrumental survey, for which we are witnessing the current trend that has led to the considerable acceleration in the possibilities of acquiring, representing, integrating, analyzing, and evaluating [2]. We are now talking of integrated hardware (tools and sensors) and software (systems and algorithms) systems, based on the rules of Euclidian geometry and epipolar, and on the more advanced techniques of modelling, image analysis, computer graphics, and computer vision [3]. All of a sudden, the scenario has become broad in terms of the offer of technological tools, more or less sophisticated, but it should not be forgotten that every single case study should be respected as a unicum. It is compulsory to identify preliminarily what the most appropriate technique of survey is, bearing in mind the goals of the survey itself in the recognition of, and in accordance with, the value of each stage and the necessity of the project.

50.2

Some Previous Activities in Using Enabling Technologies for Cultural Heritage

In the past years and still nowadays (from the 1980s until today), the author, as a researcher at Italian National Research Council (CNR), has worked intensively and continuously in the field of documentation of cultural heritage. Numerous projects have been developed, at different scales of study – territorial, urban, architectural up to the scale of detail – all faced by integrating traditional technologies with the innovative ones. Therefore, it was possible to optimize the peculiarities of each system while amplifying, through their integration, accuracy and reliability regarding the specific analyzed artifact. Different modalities and approaches have been experimented, starting from the use of manual techniques of measurement to the emergent technologies of survey such as laser scanner, multi-image-matching systems and Structure for Motion (SfM), and Unmanned Aerial Vehicles (UAVs) or drones, always integrated with each other and with the more traditional ones. Besides, the flow of acquired complex and heterogeneous data have been systematized, based on the geometries of the surveyed artifacts, in information synthesis environments, purposely designed and

50

Role of Digital Technologies, Robotics, and Augmented Realities

1469

realized, and based on GIS technologies for the systematization of flows of complex data. The case studies conducted were very different from one another in terms of the scale of intervention, importance, relevance of the analyzed artifact, and evolution of the adopted technologies. Diverse experiences have been developed over more than 20 years of collaborations with supervisory bodies, local entities, and universities. Some of these experiences – occurred in the first decade of the twenty-first century – are hereafter cited. Undoubtedly, the Roman city wall of Augusta Praetoria (Aosta, Italy) needs to be mentioned: it was entirely surveyed with digital photogrammetry technology and restored into an environment researchable with stereoscopy methodologies for the analysis and planning of the maintenance [4]. Broad documentation was established for a GIS database in the Park of Appia Antica, Rome: digital photogrammetry systems were integrated with laser scanner during the acquisition of data [5]. A similar approach of integration of digital photogrammetry with laser scanner was adopted for the Roman Theatre [6] and August’s Arch [7] as well as for the capitals of the cloister – and the cloister itself – of the Collegiata of Sant’Orso, Aosta [8]. Moreover, at Collegiata, a photogrammetry survey of the early Medieval paintings on the ceiling was conducted. All these instances were part of programs of conservation aiming at a careful analysis of the of conservation and planning of the interventions (Fig. 50.1). An image-based survey of the Balteo (second century AD), located at the Archaeological Museum of Aosta, was used to realize a virtual realization of the object. The same technologies were adopted for the frescoed spaces of the Castle of Quart to verify the removal of material in the stages of uncovering through procedures of laser ablation [9]. As part of projects of conservation and enhancement, it was then possible to test UAV systems (drone), integrated with the use of SfM methodologies on the field for some particular artifacts, in the archaeological thermal area of the Roman period at Montegrotto Terme (Padova, Veneto), Tiberio’s Villa, and the Grottos of Ulisse (Sperlonga, Lazio) [10]. A similar experiment has been conducted for the entire monumental complex of Porta Praetoria (Aosta), in order to create accurate documentation of its state of conservation [11]. Furthermore, the great number of cases concerning statuary should not be forgotten – from the Archaeological Museum of Venice to the one in Sperlonga, from the Oriental Museum at Altemps Palace in Rome, up to contemporary artworks stored in several museums (Fig. 50.2). These occasions have represented an extraordinary terrain for testing, specifically the in-depth examination of the comparison among diverse instrumental techniques. The work carried out as part of the project Asia and Mediterranean “Digital Technologies for the Confucian Temples of the Hunan Province” was highly important because it analyzed the state of conservation and the documentation of the Confucian temples (China) through the use of survey technologies and 3D modelling. The project was in partnership with the Hunan Province and under the aegis of the Beijing authorities [12].

1470

P. Salonia

Fig. 50.1 Survey experience in Aosta Anaglyph from the photogrammetric survey of the Roman city walls by laser scanning post-processing: mesh of the Arco di Augusto; post-processing data from SfM survey: point cloud of a capital of the Sant’Orso cloister; post-processing data from integrated survey SfM, UAV, and topography: Porta Praetoria point cloud (screenshot by Paolo Salonia)

These Confucian temples, built during the dynasties Tang (618–917 A.D.), Song (960–1279A.D), Ming (1368–1644 A.D.), and Qing (1644–1911 A.D.), were all severely damaged by acid rains and in need of technical interventions of monitoring and diagnostics. As part of the project, the temples were surveyed experimenting technologies of digital photogrammetry on these complex architectures [13]. More problematic were the situations where the investigation of the environmental conditions of the site represented an important risk factor for the investigator. Typically, hypogean spaces characterized by elements of cultural interest belong to this category, both for the presence of structural deficiencies with the risk of collapse and for the potential presence of radioactive elements specific to certain rock materials. In these situations, robotic systems seem to be the most suitable. So far, they have been almost exclusively adopted for interventions during emergency situations, such as post-seismic events. Between 2013 and 2016, this topic has been discussed with a European Project within the seventh Framework Programme for Research and Technological Development (Project ROVINA) [14]. The research, strongly interdisciplinary and developed in partnership with Italian and European Universities, allowed researchers to implement and extend the current knowledge on the reliability, precision, and

50

Role of Digital Technologies, Robotics, and Augmented Realities

1471

Fig. 50.2 Survey experience from UAV to SfM. On the left, post-processing data from UAV survey: Tiberio’s Villa, and the Grottos of Ulisse (Sperlonga, Lazio) point cloud (screenshot by Paolo Salonia). On the right, post-processing data from SfM: Greek female headless draped statue (Palazzo Altemps, Rome) 3D model point cloud (screenshot by Paolo Salonia)

autonomy of robotics. The main goal was to achieve a highly innovative scenario for the automatic mapping of areas of high archaeological value that are otherwise difficult to access. Further aims included the development of methods for the construction of accurately texturized 3D models, with products generated in realtime, sites of great dimensions, and with complex information and semantic annotations, to be realized through the application of innovative integrated techniques for the interpretation of visual and measurement data. In addition, the project has developed advanced techniques of autonomous and secure navigation for future experimentation within heritage sites. To this end, remote control interfaces for robots have been realized. In close partnership with the Pontifical Commission of Sacred Archaeology, two case studies were selected as emblematic situations – the Catacombs of Priscilla in Rome and those of Saint Gennaro in Naples [15–24] (Fig. 50.3).

50.3

Heritage/Knowledge/Technology. Utopias and Dystopias

Based on the experience gained and looking at the international context, despite being aware of the significant contribution brought by technological innovation, we need to question ourselves precisely on the relationship heritage/knowledge/technology in order to verify if we are really heading in the right direction.

1472

P. Salonia

Fig. 50.3 ROVINA project. Above, the robot in acquiring data (on the left Catacomba Priscilla in Rome, on the right Catacomba San Gennaro in Naples). Below left, two point clouds by laser scanning (Catacomba Priscilla sections). Bottom right, robot data acquisition control interface with point cloud production in real time (screenshot by Paolo Salonia)

Or, conversely, if a turnaround should be sought in order to increase our understanding of the extraordinary acceleration of one of the elements of this relationship, technology. In other words, we should ask ourselves if it is urgent a critical rethinking able to challenge some certainties. It is now usual practice to combine terms such as heritage, knowledge, and conservation with others, such as technologies, digitalization, 3D, and modelling – transforming the latter into the only actions able to support the entire cycle of safeguarding of cultural heritage. In this section, the author will refer primarily to those actions directly linked to the fundamental moments of conservation, enhancement, and enjoyment of heritage. The issues linked to the use of technologies to enable the understanding and communication of cultural heritage are discussed. In this regard, it is useful to briefly touch upon the meanings and definitions of the terms at stake: • Heritage: broadly and widely discussed, and too strongly ascribable to topics of legacy, it is frequently used with a superficial economic meaning; heritage will here be adopted in its full extension of cultural heritage, with the layers of

50

Role of Digital Technologies, Robotics, and Augmented Realities

1473

meanings of eco-system, and universe, the knowledge of which requires a systematic and holistic approach. • Technology: its meaning, well-known and abused, derives from the Greek tékhneloghìa, which literally can be translated as dialogue (or reasoning) above art, where art meant the ability to do until the eighteenth century. • Knowledge: the inalienable cyclic process which should accompany every human action, even more significantly in this context the action of safeguarding heritage. Today the cognitive process of a cultural asset is entirely developed within technological systems which have their own rules and which are created to achieve goals not necessarily related to heritage interpretation. In one of his essays, the English anthropologist, Jack Rankine Goody, provides the following definition of technology: “codified ways of deliberately manipulating the environment to achieve some material objective” [25]. This definition well matches Marshall McLuhan’s representation of the evolution of technology as a progressive extension of the human body. McLuhan, in fact, defines technology as “a way to translate one system of knowledge into another” [26]. Therefore, it is important to set the goal of our knowledge. If a conservation intervention is required, the goal of the process of knowledge is to characterize the heritage element in its morphometric (knowledge of quantitative and spatial dimension), historical, material, and phenomenological (chronological and qualitative knowledge) entirety. Based on the results of this process, it will subsequently be possible to plan the intervention, execute it, and finally monitor the behavior ex-post on the same cognitive basis. In this regard, Stefano Brusaporci writes that in the “analysis, study, documentation, and representation of architectural goods, the digital dimension has to face the reality of the architecture, synthesis of spaces, surfaces, volumes, materials and technologies (author’s note: but also of phenomenologies, and manifestations of kinetic material decay), a result of the processes of modification and stratified testimonies of the events and architectural cultures which have occurred over time” [27]. However, the practice is heading in the direction of a (correct) systemic vision aimed at producing economies of scale. In this way, the same data of knowledge derived from the use of technologies are adopted for other moments, closely related to the enjoyment of heritage. However, the elaboration of these data – oriented to communication strategies addressed to diverse types of customers, and in a globalized acceptation of the concept of enhancement – are almost always addressed only to a touristic use. Therefore, it is now routine to decimate 3D models realized for conservation projects, decreasing their resolutions, making them immediately available for the communication of heritage as a metaphor for the integrations of other contents, and for its realization in immersive spaces. Cyclic and implementable modalities of management of heritage are guaranteed on the same cognitive base. Within this procedural schema, the survey (still) plays a central role, since the measurement of sizes and forms is the scientific way to construct a model which corresponds to reality.

1474

P. Salonia

Several are the examples that testify how the evolution in the field of technologies applied to the survey of heritage has increased significantly in terms of accuracy and reliability of the measurement, restitution, and management of the data acquired. However, not always these applied technologies have been accompanied by a simplification of the procedures and a lowering of costs. As previously observed, in a very short period of time, we have changed over from the traditional systems of classic stereoscopic photogrammetry to SfM methods, from the simple topographic measures to the use of laser scanner which restores a huge quantity of spatially measured points. In relation to these technologies, image-based and laser scanning, it should be noted that, in the light of the ethical decisions in the procedures to adopt, it is preferable to adopt the former since significantly low cost but with almost the same reliability of results. Undoubtedly, the introduction of these technological tools has implied partial procedural reversal of the traditional survey: the stage of measurement is substituted by the production of a sort of digital cast discretized of the entire artifact (the points cloud), generated in semi-independent modalities. These procedures have substantially diminished the direct involvement of the operator, whose critical thinking comes into play primarily in the phase of post-processing of the restitution of data related to the artifact. Therefore, it is a duty of the modeller to evaluate the possible redundancy of data while critically reconstructing the idea behind the real artifact to represent it as its alias. The person in charge of the survey no longer operates in direct contact with the real object but in a situation of alienation, where the risk of introducing subjective elements of interpretation and evaluation is very high. The adoption of technologies significantly multiplies our abilities of measurement to limits that would be unthinkable if we had to rely only on our perceptive senses, and manual skills. From this awareness derives a first reflection on how the technological mediation interposes itself between us, subjects of knowledge, and the artifact to know. Frequently, in fact, we witness a decrease in the attention paid to real objects of study, entirely substituted with an increased attention toward the instrument of measurement. This matter was already discussed by Walter Benjamin when introducing a dialogue on photography, he writes: “For the first time in the process of pictorial reproduction, photography freed the hand of the most important artistic functions which henceforth devolved only upon the eye looking into a lens. Since the eye perceives more swiftly than the hand can draw, the process of pictorial reproduction was accelerated so enormously that it could keep pace with speech” [28]. W. Benjamin refers to the “reproducibility” of a work of art, while here we refer to the measurement of sizes for the 3D modelling of reality, anyway a form of reproduction of reality. This argument works equally well if we replace the eye looking through a lens with the laser ray of a scanner which restores spatial coordinates. Moreover, once the numerical model which reproduces the reality with a higher degree of reliability has been obtained, this model can generate a faithful copy

50

Role of Digital Technologies, Robotics, and Augmented Realities

1475

Fig. 50.4 Rome, via Condotti, showing 3D copies of cultural heritage: experiential performance aimed at cultural tourism or high fashion merchant promotion? (Photo P. Salonia)

simply by processing the data through a 3D printer. It is already possible to find reproductions of the Nike of Samothrace or the Venus de Milo at a 1:1 scale, in cardboard, or some other poor material, in knowing shops selling post-modern furniture (Fig. 50.4). This argument becomes of further relevance when considering the conceptual and intellectual breadth of the survey, understood as a complex operation that requires the ability to catch the unicum the studied artifact represents due to its unique features of authenticity, tangible and intangible meanings, and values. The Right to Heritage for all mankind derives from these values. W. Benjamin writes: “The authenticity of a thing is the essence of all that is transmissible from its beginning, ranging from its substantive duration to its testimony to the history which it has experienced. . . And what is really jeopardized when the historical testimony is affected is the authority of the object” [28]. For several years, the topic of authenticity somehow codified with the Declaration of Nara (1994) [29] has been, and continues to be, at the center of scientific debate. In this time of digital revolution when the reproducibility of reality has assumed strongly innovative features, especially in the field of anthropologic and cognitive sciences, authenticity is often seen as a direct consequence of the experience the public lives in front of a work of art [30]: authenticity has no longer a specific, material, and immaterial connotation, innate to the object itself because of its specific attributes. Clearly, this type of conceptualization introduces possible criticisms of the

1476

P. Salonia

risk that an excessive offer of heritage experiences can lead to a commodification of heritage and thus to the subsequent loss of authenticity [31]. Other authors define these new forms of interactions between people – either an individual visitor, the public, or a community – and cultural heritage as a performance, and such meetings as contemporary authenticity [32]. In this sense, these processes would be able to transfer the aura of the artifact, since it is enhanced and enjoyed in innovative modalities, to the performance which would generate new meanings and values. In this way a debatable cultural value (and of guidance for the cultural economy) is attributed to visitor experiences which become the only forms of heritage enhancement. Furthermore, concepts such as embodiment [33, 34] are introduced to underline the role of immersion and presence that virtual, and augmented reality environments – and even more mixed reality – guarantees for a better knowledge and understanding of the past by visitors immersed, or better embodied, in the virtual reconstruction of the past [35–40]. Given these positions, the primacy of the Venice Charter [41] and the Theory of Restoration of Cesare Brandi [42] should be remembered as the foundations for every conservation intervention in respect to the authenticity of the artifact to safeguard, with regard to both the material aspects and the cultural testimony. Several are the considerations that could be put forward as a critical analysis to this type of approaches. It should here be mentioned the excessive subjective interpretation that can expose to temptation the operator, who is often defined as technological authority for this reason. In fact, during the intervention of virtual re-construction of the real acquired in points cloud, the operator may be led to make the visitor experience more exciting by enriching the objective reality reconstructed with the virtual subjectivity. More broadly, and to be more rigorous, it is sufficient to observe the absence of a protocol, shared and prescriptive, which defines a set of measurements and observations able to quantify the level of correspondence between the real artifact and its digital surrogate. The quality of a result of visualization can be quantitatively measured (number of pixels, density of the points cloud, number of scanning made, environmental conditions, etc.), but always delegating to the operator the choice of the threshold between the authentic and the reproduction, based on the individual level of interpretation and the available documentation. Therefore, this not always agrees with the sharp observation of Leon Battista Alberti who, when underlining the differences between painters and architects in the second book of De Re Aedificatoria, writes: “. . .as someone who wants that his work is not judged according to illusionary appearances, but check exactly against controllable measures. . .” [43]. The same criteria, which underpin the London Charter for the computer-based visualization of Cultural Heritage (2009) [44, 45], concern the morphometric characters and the aspect of artifacts, but they do not exhaust all the categories of characteristics of the objects themselves. For example, among the latter there are those characteristics measurable with the four senses besides the sight, or acquired

50

Role of Digital Technologies, Robotics, and Augmented Realities

1477

with other methods also invasive such as the chemical-physical characteristics of the constitutive material, up to include the stock of evaluations on the kinetics of decay. This argument leads us to affirm that, probably, the objectives of the 3D reconstruction and representation should be defined with extreme clarity. It should be inconvertibly traced the difference between an operation aimed at the analysis and scientific documentation and an operation aimed conversely at the creation of virtual environments as part of a museum experience or simply a touristic offer. The aims oriented at the operations of scientific anastylosis, or faithful reconstruction of parts should be separated from those oriented at various performances which stimulate visitors’ feelings running the risk that visitors assimilate a ready-made knowledge. This way it would also be possible to return full meaning to terms such as aura and authenticity as they are used, perhaps, by W. Benjamin and, surely, in the Declaration of Nara. By now we run the risk to pay more attention to the muscularity of technology and to the spectacularity of the 3D model than the real artifact of which that model represents the millimetric reproduction and the morphometric basis of knowledge. This is a risk of which we need to be aware of, also for orienting at the right dimension the use of emergent technologies in the operations of measurements. The topic, several times discussed, is that to firmly maintain the reason of the ultimate purpose (the knowledge for the preservation of the asset) and to not substitute this with the means (the technology to construct, use, and disseminate such knowledge). Unfortunately, today it often happens the contrary when we neglect the true objective of our actions so we are under the spell of the model which technology allows us to realize. Therefore, at the end our effort focuses on the research of a millimetric perfection and of image which only in modest measure represents the true cognitive demand of the project of conservation. Nor the correspondence of such precision to the real object is verified by applying procedures and standards of control universally agreed. In a sort of fideistic approach toward the technological tool (where has the Error Theory gone?), this becomes, in its infallibility, a true technological authority, and the true and the only ultimate purpose toward which our attention and our speculative effort are diverted. Furthermore, this same effort is often overly dimensioned compared to the real necessity of knowledge, and the data gathered are, for a big percentage, redundant, and will probably never be used. As underlined by Richard Saul Wurman, without doubts we are the producers of big data, but so far we are a lot lacking in big understanding [46]. In view of the issues identified, although without a detailed critical analysis of the topic able to engage not only the close academic environments but also the masses of investigators on the ground, we witness an uncontrollable process of digitalization of heritage in different global contexts and at all levels, with an exorbitant number of public, private, and public-private projects. Several are the initiatives supported by states at a global level, and it is possible to remember, compiling an absolutely not exhaustive list, the project Europeana for the

1478

P. Salonia

systematization of Digital Heritage, which concerns all types of heritage, including archives, image libraries, and so on and so forth [47]. It should not be forgotten Arches Project, an international collaboration which includes the Getty Conservation Institute and the World Monument Fund, alongside with universities and research centers of different countries, with the aim of an informative geo-spatial and open source system which represents a digital archive for the web management of material cultural heritage [48]. Similarly, it is also worth mentioning the project Cyark 17 of the homonymous Technology Centre [49], or the project of Mosul, realized thanks to the collaboration of The Economist with the no-profit group Rekrei, and which has begun after the destruction of significative heritage elements in the North of Iraq [50]. From these limited references, it should not be omitted the Herculaneum Conservation Project, sponsored in the past years by Hewlett Packard and the Swedish Pompei Project, which has been promoting the digitalization of different insulae in the site of Pompei since 2010 [51]. In addition, all around the world many investigators are daily engaged in activities of digitalization. These investigators act on behalf of university, research centers, supervisory bodies, administrations, local, national, international institutions, and private non-profit organizations. Images are produced, laser rays are shot, and points cloud thickened, but, in proportion, fewer restoration sites are opening. An incredible plurality of voices, dialects, and languages (in a technological sense) reigns supreme in this chaotic and strongly de-regularized scenario, where there are no universally standardized references to which practices (more or less good?), modalities, and necessary minimum requirements should be referred to. What are the indexes and parameters to verify the accuracy and reliability of the produced data? What are the certain scenarios where it is possible not only to verify but also compare different data? And what are the criteria to make these data comparable also in relation to the above-discussed activities of monitoring? What are the administrative bodies and the authorities to contact? Several projects, perhaps too many, are undergoing on this topic, but it is not possible to affirm with undisputable certainty that universally agreed protocols exist yet. Despite the principles for the recording of monuments, groups of buildings and sites, ratified by the 11th ICOMOS General Assembly in Sofia (October 1996) [52], the London Charter for the Use of 3D Visualisation in the Research and Communication of Cultural Heritage (2009) [44, 45], the Manifest of Digital Humanities, Paris, Thatcamp (18–19 Maggio 2010) [53], and the Principles of the Seville Charter - Spanish Society of Virtual Archaeology (SEAV) (2008–2012) [54], the international community is still too far from finding a single voice. Not surprisingly, also the 26th Biennial Symposium of CIPA “Digital Workflows for Conservation” – which was aware of the difficulties to identify standards of practice in the field of cultural heritage – attempted to explore and compare different approaches, and to evaluate the results so far achieved in the production and management of digital data for the conservation of cultural heritage [55]. In one word, in this (unintentional?) process of change from the real into the virtual, the concept of safeguarding of material heritage, which imposes processes of

50

Role of Digital Technologies, Robotics, and Augmented Realities

1479

knowledge and conservation, is now often confused primarily, if not exclusively, as a digitalization of heritage. Striking evidence of this is the aim of most of the European programs in the last years, and still nowadays. In these projects – in view of the declarations of principle on the irreplaceable role of culture and cultural heritage for the education, enhancement of people’s identity, and growth of societies in an inclusive, innovative, and reflective sense – types of research based exclusively on digital content have been favored. In practice, the promotion of impressive process of digitalization, from archive to specialist lexicons, from monuments to sites, has been constantly supported. As examples, it should be remembered: Michael-Multilingual Inventory of Cultural Heritage in Europe [56]; EPOCH (European Research Network on Excellence in Processing Open Cultural Heritage) [57], 3D-ICONS (increasing the critical mass of engaging 3D content available to Europeana’s users) [58], CHARISMA (Cultural Heritage Advanced Research Infrastructure SMA) [59], DARIAH (The Digital Research Infrastructure for the Arts and Humanities) [60], JPI (Joint Programming Initiative on Cultural Heritage) [61], VII Framework Programme [62], Horizon 2020 [63], and ERIHS (the European Research Infrastructure for Heritage Science) that supports research on heritage interpretation, preservation, documentation, and management [64]. The different calls of these research funding programs refer more to the field of hard science than the one of humanities. The evidence for this is that, in the 2-year period (2014–2015), in Horizon 2020 – Challenge 6 [65], the theme of cultural heritage is cross-cutting but based on disciplines aiming at the advancement of knowledge on 3D modelling, net-working, and digitizing. For example, the topic “Advanced 3D modelling for accessing and understanding European cultural assets” is allocated a budget of 14 million euro, while the “European research infrastructures for restoration and conservation of cultural heritage” only c. seven million euro. Therefore, there is a need to question how much our critical-conceptual apparatus is adequately developed to control the smart use of technologies which speed up, improve, iper-realistically visualize a reality of artifacts to safeguard orientated to their unstoppable digitalization. Meanwhile, perhaps, we do not realize the progressive material loss of these artifacts, where the ultimate purpose of the conservation and enhancement process is to pass on to future generations the materiality of heritage. This material testimony will allow the present and future understanding, dissemination, and sharing of the meaning of cultural heritage as an unalienable human right. This testimony and its sure accessibility will be the further and unchangeable guarantee for the realization of a society based on identity, inclusivity, solidarity, and equality. Imagine to disjoint the diverse phases of approach to an asset, and begin to analyze those basic phases concerning the methodologies for the knowledge of the Geometric Consistency and Shape, the Material Consistency and State of Conservation, and finally the Historic-Critical Aspects. Considering the current status of art, only for these three phases, it is possible to speak in terms of useful enabling technologies. In fact, in these phases, we have to adopt actions of morphometric characterization and analysis of historiographical

1480

P. Salonia

Fig. 50.5 Ideal pyramid of the approach to an asset: from the knowledge to the perception of identity value, from analysis to the Right to Heritage. Where is the role of technologies? (Drawing P. Salonia)

documents. In practice, we have to deal with quantity and quality, measurement, and description, for the conservation, communication, and enjoyment – actions for which, as above discussed, technologies have become specialized in and progressively innovative. However, when we want to face the subsequent aspects – typically inherent to meanings and values and, even more, to the Right to Heritage – it is easy to notice that the technological contribution is substantially missing as a decisive tool for providing answers to the so-called historic question (Fig. 50.5). To what extent, do enabling technologies support us in understanding meanings and values? How do they guarantee the Right to Heritage, excluding the mere media power of digital communication? In digital communication, the means to convey the message is undoubtedly strengthened, but the content of the message itself remains substantially unchanged, lacking of any innovation and in-depth analysis. Moreover, the specificity of our current situation, and even more of our sector, assists to the consolidation of a more divaricated gap between the technological development and the delay in the disciplinary debate for the affirmation of a diverse use modality of such technologies. From one side, this technological development autonomously follows an exponential growth according to its own rules which are different from the diverse areas where technologies can be applied. From the other side, it has not been defined and had yet modalities based on critical awareness able to question the real and always necessary validity of that amount, surely fascinating but sometimes unmanageable, of complex and extraordinary (but how much essential?) big data. Furthermore, we well know that the technology we adopt was born in other contexts and for satisfying other needs. Even more, we are aware of how the field of

50

Role of Digital Technologies, Robotics, and Augmented Realities

1481

conservation and enhancement is considered, within market logics, a niche sector to refer to only to recycle technologies. The technological development has an autonomous life generated and developed from other factors and rules, addressing the market toward new needs and necessities. Therefore, others are the applied fields which determine choices and move huge resources able to address the industrial strategic policies. The technologies industrialized according to the original purpose (to solve specific issues linked to the knowledge for the conservation project or the identification of new forms of enhancement and enjoyment of heritage) are very rare. The applications – born for spreading diverse forms of enjoyment of heritage on devices, smartphones, and tablets – are presented as “enhancement,” since the use of virtual, augmented reality, today mixed reality, guarantees a major dissemination (without doubts) and a major wealth (they could), besides a more in-depth knowledge (are we sure about it?) of the visited site or monument. However, these devices, undoubtedly, are born independently from their potential application to cultural heritage, and the related technology is often aimed to raise needs before not perceived. Then, it becomes useful to wonder whether the modality – with which the cultural asset is known and communicated – is inspired by specific considerations on the meaning of that asset. And, furthermore, how is the knowledge of a cultural asset conveyed in order to build identity awareness and wealth in the citizen? It is easy to recognize that undoubtedly the means influences the content of the message. Therefore, once more, McLuhan is right. The discussed problem presents cultural and socio-economic characteristics where the so-called digital divide – which is used as a more general form of technological divide – poses (digital) technologies for heritage as a new form of colonization. Examples of this are the countless cultural missions (archaeological excavation, documentation of sites, etc.) carried out by opulent countries in territories, formerly the cradle of civilization, but today dramatically forced into situations of backwardness, poverty, and hunger, often even tragic theaters of others’ wars. After the display of performing technologies, big collections of data, and discoveries, what is left to the indigenous communities in terms of social enhancement built on the values of a rediscovered identity? What is the cultural, civic, legacy of the progress? What is the economic growth? Rich countries themselves remain then totally passive and silent in front of criminal devastations such as the one of the Bamiyan Buddhas or the archaeological evidences of Palmyra [66]. These countries more easily propose, however far from those theaters, post-operam interventions (where operam is the terrorists’ destructive fury), constructed through 3D printers or advanced holograms for the reconstruction of such evidences. These are material (innovative and technological 3D copies) or virtual (ectoplasms or will- o’-the-wisp of what once was real) reconstructions. It could be discussed here the wide issue of the military devastation of cultural heritage as effective and terrible weapon to target the enemy’s higher and deeper values as part of his cultural identity. This offence represents the further and bigger

1482

P. Salonia

demonstration of what is – or should be – the role of cultural heritage in building social capacity and cohesion. To this picture, per itself dramatic, it should be added that often different public bodies – in charge of the safeguarding of heritage – commission to private third parties, possessors of advanced technologies, and skilled merchants to produce enormous quantities of technological scripts. These products are very advanced as well as vain and unutilized since they are destined to become, in a very short period of time, obsolete digital materials. More easily, in the same public organizations, it occurs that the hardware infrastructures for the reading of these products result to be fatally obsolete. The International Council on Monuments and Sites (ICOMOS) deals regularly with the urgency of a close debate on the issue here discussed. In the Florence Declaration on Heritage and Landscape as Human Values, adopted during the 18th ICOMOS General Assembly in Florence (2014) [67], Action 5 identifies the following fundamental principles for a responsible use of technologies: 5.1 Cultural heritage objectives need to drive the development of emerging innovative tools, not vice-versa. 5.2 It is necessary to promote new technologies that are accessible and inclusive for shared cultural growth. 5.3 It is urgent to facilitate collaborative standardization and simplification of procedures and tools. Surely new opportunities are emerging for the systematization of huge amount of information, for the digitalization of cultural heritage, and the use of social media. The final goal is to support the control, planning, organization, management, interpretation, and monitoring of the conservation actions but also to identify new forms of enhancement and enjoyment. This should be done without forgetting the new numerous factors at risk, from climate change to globalization, from security to the demographic growth. To rule this not only technological but also primarily cultural change, fundamental principles should be able to guarantee respect for diversities, a systematic and holistic approach to the conservation process, the continuum cycle of knowledge, project, intervention, and respect for the local and traditional knowledge. The topic of meanings, values, and the Right to Heritage has been introduced in the previous paragraphs. In the next two sections, it will be useful to examine what will be, in relation to this topic, the possible future scenarios according to the analysis of the current situation at the end of the second decade of the twenty-first century.

50.4

Technological World and Future

The twenty-first century made its own, giving them a further impulse and declension, those values which at the end of the 1800s had been largely theorized by the western cultures. Especially in Europe, this ferment has been particularly active following the

50

Role of Digital Technologies, Robotics, and Augmented Realities

1483

evolution of a way of thinking which had its roots in the Italian fifteenth century, and which was not minimally affected by the subsequent historical events and the imposition of new models of society and development introduced by the Industrial Revolution. On the other side, the steady development of this thinking followed, especially democratizing it, a historic/cultural process that was believed to play a paramount role in the progress of mankind. In particular, in the post-military years but increasingly in the following decades, the atmosphere of general attention toward the meaning of cultural heritage in modern society determined the rise of important international bodies, which became soon promoters of confrontations, debates, studies, and research. It is an example the establishment of ICOMOS [68] a year after the Venice Charter of the 1964. It is evident how a cultural fragment of such entity managed to raise in the public opinion a new and diverse awareness of the importance of the meaning and identity value of cultural heritage, despite people were strongly projected into the oxymoron of a vision of an illimited progress and endless growth. The specific nature of universal asset will be attributed to cultural heritage as a right of all mankind (UNESCO 1972 Convention Concerning the Protection of the World Cultural and Natural Heritage [69]). Simultaneously, cultural heritage will be seen as historic evidence of communities’ specific identity. For example, in Italy we can bound the core of this concept of identity in the final document of the Commission Franceschini which, already in 1966, defined cultural heritage “everything that represents material evidence of cultural value” [70], principles definitively embedded in the Italian Codex of Cultural Heritage and Landscape (2004) [71]. At the end of the second decade of this twenty-first century, future scenarios appear to be difficult to define and trace with certainty. The uncertainty of the future, especially the absence of clarity on the principles and values that will constitute the bearing structure of the future, determines a central aspect of the deep crisis, a further result of globalization, which the entire world is going through while wrapped in an invisible but pervasive and noisy infosphere. In this complex scenario, it becomes paramount to urgently wonder what will be the future of heritage and historic cities, and whether their identity value will still be acknowledged? More precisely, it is important to understand to what extent the coming society will want (and will be able to) to inherit the future left, for better or worse, by the previous century and to re-establish a new model of development built on the central role of cultural heritage and culture, recognizing in these two elements its inner values. Today it seems that the fundamental terms at stake are registering a violent and unstoppable reversal of trend, hardly decipherable, that appears entirely dichotomous as regards a very recent past. In fact, both the first and the second Industrial Revolution were marked by their in-depth economic and social movements, the unavoidable cultural transformations, and especially the tragic military consequences derived from this period. Similarly, in the years of reconstruction, the third Industrial Revolution, subsequent to the Second World War, imprinted profound changes in the socio-economic apparatus at global level.

1484

P. Salonia

However, the so-called digital revolution, sped up in the last decades, introduces elements of such great impact in people’s everyday life that it constitutes a real solution of continuity with the historic process. It is possible to affirm that the digital revolution clearly traces a limes between the past and the future. It is not only about the incomparability of the speed reached by the technological innovation that, as a new form of power, is different but also because it possesses now logics and paradigms (apparently) autonomous and that escape the known arguments on the traditional process of production. This innovation, erected as a form of power, undoubtedly is integrated into the financial power which has violently taken the place of traditional economies and politics, in an ambiguous relationship that makes difficult to understand which of the two powers rules the other. Together, technology and finance, clamp the globalized world in a vice, with the same meaning of the term coined by McLuhan, with extraordinary second sight, in the 1960s of the last century, at the beginning of the electronic revolution. Technology produces hardware, software, means and machines, but cannot produce structured cultural models and supplemental ideologies (it may be said: luckily!). Currently technology defines and forces new behavioral models and stimulates ephemeral “needs” produced outside our consciences. The effects are constantly in front of us, and we are witnessing a progressive anthropological transformation, the results of which are currently unpredictable. This concerns not only the so-called Generation Z (those born between 1995 and 2010), but also the generation of the “heroic” Millennials, or digital native (1980– 2000). Finally, as a consequence, the anthropological transformation concerns also extensive areas of fathers’ previous generations who, at the very least, have delegated many of their educative tasks to social and mass media. It is a transformation that goes beyond a reductive discourse limited to the technological prosthesis, the extraordinary performing devices that fill our houses and which today we take advantage of also in different social uses. Partly unaware, we have algorithmic powers which were prerogative only of big computation, and research centers until few years ago. At the University of San Diego, a study has been conducted on the psychological impact that new devices, smartphones, and tables are having on the development of children and adolescents born in a world already ruled by these devices. The study used a sample of 11 million young people. In the article, Twenge writes: “Over the past decade, adolescents and young adults have spent an increasing amount of time using. . .electronic devices. . .U.S. 17- and 18-year-olds spent 6 hr a day texting, online, and on social media. . . More frequent digital-media use may also displace time spent on activities more beneficial for happiness and mental health, including face-to-face social interaction. . .adolescents and young adults who spent more time using digital media reported lower psychological well-being, including more stress and psychological problems. . .more feelings of loneliness. . . more difficulty making friends” [72]. Moreover, while investigating the future potentialities of Internet of Things (IoT) and the children’s relationship with digital devices for learning, interaction, and games, a team of researchers of MIT Media Lab underlined the potentialities of

50

Role of Digital Technologies, Robotics, and Augmented Realities

1485

Internet of Games (IoG) stating that “. . .connected toys will become the norm and this will lead to children (note of the author: children who do not know how to catch the real emotions of their real fellows) expecting toys to not only respond to their commands, but to show emotional intelligence of their own” [73]. These studies and research are undoubtedly useful to the games industry for the children of the new millennium, but they are also more oriented to the progressive transformation of many of the human activities into a continuous and unstoppable gamification of the real world. Michael Merzenich, professor emeritus at the University of California in San Francisco, one of the most well-known neuroscientists in the world, warns us: “we think that to remove efforts and assign them to a devise is always an advantage, we forget though that every time we assign to a machine a human function we are removing something from our life and our brain. Technology is changing us and it is changing us in the brain. It is amazing to have access to a multiplicity of information but if do not exercise the logic and the reason we are in trouble” [74]. Today we confuse what is user-friendly with what is not easy to learn: we have direct access to an endless amount of information, apparently for free, but we have not yet developed a logic, a culture to better use this mountain of notions and extract from them the right conclusions. Merzenich researched the hypothetical advantages of a sudden turnaround from complexity (indefeasible feature of our life) to simplicity, meaning simplification. For example, he imagined an electronic device that, once heard our question, can answer us in few seconds giving the solution for every topic. In doing this, he explored what would be the practical effects on our brain if people live in a world where technology simplifies every aspect of their life [74]. This is exactly what is happing right now. If we think of Alexa, the multitasking smart assistant to which people entirely rely for the management of their days, from the alarm clock in the morning with diffusion of classical music to the turning on of the lights in rooms, from the sending of messages to the answers necessary for the development of children’s homework. While adults have already lived part of their life in contexts with no technology, something that makes them able to partly limit the damages, the youngest have no type of protection. Every life experience for them has begun in a world already at the highest technological density, and this represents a serious problem. From the analysis of data coming from 239 countries, the Global Digital Report revealed that “global internet users have now passed the 4 billion mark. Well over half of the world’s population now uses the internet” [75]. The smartphone is now a device which we cannot easily separate from. In the world, there are more SIM as regards the 7 billion people. Neuroscientists noticed that people are testing their brains to think by jumping from one idea to another as they pass from one link to another. People are no longer training their brains to think in-depth, to ask themselves detailed questions, to be thoughtful: they do not stop anymore to evaluate matters. People limit themselves to simply search online at high speed, and the consequence is that they are becoming only reactive creatures.

1486

P. Salonia

If we think about the new impressive frontiers open by Artificial Intelligence (AI), it is not exaggerated to affirm that there is a new ethical matter, and it is useful to wonder whether new technologies risk to determine the end of the Humanism or, alternatively, to establish a new one. Undoubtedly, the virtuous circuit mankind-nature-technology is being interrupted, with the introduction of new paradigms through which virtual reality, or even augmented reality, overlaps – up to take the place in our deep perception – with our experience of the real world in a new and unknown ecosystem. This happens without being the necessary interpretative instruments of our senses (of our new and diverse relationship with the reality) addressed, and without being the new systems and meanings being defined. As observed by Umberto Galimberti, “We continue to think of the technique as an instrument at our disposal, while the technique has become the environment that surrounds us and shapes us according to the rules of the reason which – measuring only against the criteria of functionality and efficiency – do not hesitate to subordinate human needs to the needs of the technical device. Unaware, we still move with the typical traits of the ‘pre-technology’ man, who acted in view of goals inscribed in a horizon of senses, with ideological baggage and a stock of feelings in which he used to recognise himself. . .The technique in fact can mark that point, absolutely new of history and maybe irreversible, where the question is no longer ‘what can we do with the technique?’ but ‘what can the technique make of us?” [76].

50.5

Cultural Heritage, Identity Values, Memory in the TwentyFirst Century

Certainly, what above described represents a probably forced scenario but not for this reason unlikely to happen. It will also be useful to discuss how the discipline of conservation, from safeguarding to enhancement, needs to face the emerging topics of the twenty-first century, in order to rethink the theory and the practice of its becoming and development in dynamic forms. Cultural assets, as broadly sustained, represent shuttlecock of civil progress for the re-foundation of a society that does not lose its identity and its values. Facing the topic of heritage safeguarding, within this hard-to-define scenario, we need to ask ourselves whether we will still share these cultural and identity values, as well as our capacity of recognizing the ethical meaning and value of cultural heritage. In order words, we need to ask ourselves if our baggage of thought, knowledge, theoretical speculations, and scientific elaborations – which is at the core of our today awareness – is otherwise condemned to be slowly but undoubtedly forgotten. Current affairs lead us to foresee a future where the technological prosthesis – the medium always more active – will be the one to decide what the message should look like, and what the modality of presentation to the “human” recipient – who has become an (almost) passive object – will be. These are very central topics, which go from historic centers to the scale of the single artifact which should be preserved as much as possible in reality (to preserve

50

Role of Digital Technologies, Robotics, and Augmented Realities

1487

material and meanings) rather than being digitalized (to create a digital alias). This would prevent historic centers from being commonly and hastily defined as smart historic cities. This definition represents a trendy new locution which in reality risks to appear as a rush label which does not deepen any cultural advantage besides the superficially informative one. In fact, what continuity is guaranteed between the depth of the transformations settled over centuries and, for example, the use of QR Code on historic buildings to orientate tourist flows with various information (often banal) displayed on different devices? What are the real contents aimed at raising the quality of citizens’ life in historic cities made smart? How much invasive could the interventions of wiring be? How much efficient could the wireless ones on historic and monumental buildings and on historic urban fabrics be? Finally, what officially parametrized resources and checks will be used? More optimistically, Carlo Ratti sustains: “the historic cities have never been able to adapt to the technologies of the previous century, heavy, invasive, and not compatible with our cultural heritage and the urban morphology. . .On the contrary, the technologies IoT (Internet of Things) can easily adapt to the historic urban fabric.” According to Ratti, in fact, “while we explore the consequences of IoT which do not require infrastructure, we should start working on one city as it is. . .this is very different from what happened in the previous century. . .” considering that “. . .Smart City is nothing than the result of a wide technological phenomenon. . .if we play well, these transformations can help in making urban areas more sustainable and producing new forms of citizen participation. . .” [77]. Going further, he almost depicts oneiric visions according to which the smart city will be a “. . .playful city. . .the dream of a New Babylon. . .a society of total automation, the necessity of work is substituted by a nomadic life of creative play, a modern return to Eden. . .(where) ‘homo ludens’, once day. . .will get rid of work, he will not have to make art, because he will be able to be creative in the practice of his daily life” [77]. If we look at the current affairs of contemporary cities, in particular, metropolitan areas with strong historic features in their central urban fabric, we can affirm that these cities represent the metaphoric and physical place where complex phenomenologies take place and develop. These phenomenologies are interactive, systemic, and require to be dealt with using a holistic approach and technological innovation while being aware and respectful of cultural values. Alternatively, the crisis of globalized cities – especially of those defined with awful locution art cities – derives from the contradictions of the current model of development. There is an incapacity of finding, in the awareness of one’s identity, the necessary boost for a change of paradigm able to base on culture the vision of a different perspective of growth and sustainable development of the cities, landscapes, and societies. Reduced to touristic-archaeological evidence, historic cities have seen their civil character being mortified, increasingly transformed into a city-postcard, reduced to an untouchable freeze-frame which collocation is hard to be determined. In this regard, Salvatore Settis writes “the souvenir photos taken by tourists, the quantity of which is more important than the quality, certifies not the cultural

1488

P. Salonia

curiosity, but the ritual presence of the tourist, they do not archive the memory, but they substitute the gaze and the memory” [78]. The emotional performance – discussed in the previous pages as an innovative concept and justification of an apparent virtual authenticity – risks, too often, to be reduced to a sterile, repetitive practice. We limit ourselves to conduct a compulsory ritual which nothing has to do with the introjection of a meaning, a value, or a memory. Alternatively, to this monothematic cultural offer in the real world, will people be more interested in walking among the ruins of the Roman Forum in Rome, or in navigating through it wearing an Oculus visor which allows to live an experiential performance within a mixed reality environment of a serious game? Moreover, in the absurd mercantile competition which is gradually leading to the death of our historic centers, the vulgar market of touristification produces products of virtual, augmented, or mixed reality which have nothing in common with what the research in the cultural sector analyses and studies. This unstoppable homologation condemns the true nature of historic cities to mortal immobility, in a unique cliché: it contrasts and suppresses these places to a static vision entirely in contrast with the concept of dynamic mutability, shape, and substance of the cultural landscape of a place. In addition, this bulimic consumption determines the “desertification” of cultural values of historic cities, causing paradoxically the definitive loss of the attractiveness which draws visitors. Italo Calvino writes: “forced to stay immovable and always the same to be better remembered, Zora languished, disintegrated, and disappeared. The Earth has forgotten her. . .” [79]. This mercification, passed off as democratization of culture, presupposes the imposition of a false and standardized version of what past, culture, and diversity we need to be fed with. From here, the use and indiscriminate abuse of monuments – places deprived of their genius loci – landscapes, cultural lands. Other monuments, places, landscapes, and lands – declassified to a lower category in a scale of purely monetary values – are left totally abandoned, forgotten, and progressively impoverished. It should be here introduced a clear distinction among safeguarding, conservation, and heritagelization. The latter derives from a process of selection to respond to the cultural and identity needs of the current dominant society. This heritagelization has as its own collateral effect the so-called dissonant heritage [80], meaning the mechanisms which may de-heritagelize broad quantity of assets and lands based on the acknowledgment or disavowal of their spirit of place. This occurs when a part of a community (local or global) – in harmony or in contrast with the evaluations of those who decide – operate a selection (also unconsciously) based on the dominant historic-cultural trend, when not influenced by the big lobbies of the touristic circuits. In essence, we can affirm a parallelism: as the processes of heritagelisation are established and managed by the emergent élites of societies, in the same exclusive

50

Role of Digital Technologies, Robotics, and Augmented Realities

1489

way the technologies and the modalities to communicate and enhance assets recognized as heritage are chosen and used. If this is the reality, and unarguably it is, the discrepancy between it and the solemn appeals ratified in global declarations appears in all its evidence. From the reading of these documents, the loss of the Right to the City as a common heritage and the Right to Cultural Heritage and Landscape are serious attacks to human rights. To mention just a few of these declarations: • The Faro Convention (27 October 2005) establishes the fundamental principles for raising the awareness of cultural heritage in Europe, and the contribution heritage brings to the wealth-being: cultural heritage and landscape (the possibility to enjoy them) as universal values and rights and shuttlecock of inclusive and informed growth [81]. • The Quebec Declaration (2008) on the safeguarding and enhancement of spirit of place [82]. • The definition of Historic Urban Landscape – HUL (2001) [83]. • The Hangzhou Declaration (2013) establishes Culture as key to sustainable development [84]. • The Florence Declaration (4 October 2014) acknowledges “the role of culture as an enabler and driver of sustainable development and which requested that culture be given due consideration in the post-2015 development agenda” [85]. • The United Nations Sustainable Development Goals (SDGs), objective 11–Sustainable Cities and Communities of the Agenda 2030 which UNESCO has adopted in 2015 during the 70th Session of the United Nations General Assembly. According to UNESCO, the document represents “a universal, ambitious, sustainable development agenda, an agenda ‘of the people, by the people and for the people’” [86]. • The EU Urban Agenda of the European Commission (2016) which focuses on three top priorities: smart, green, and inclusive cities [87]. • The New Urban Agenda (2016), adopted by the United Nations during the Conference on Housing and Sustainable Urban Development at Quinto, Ecuador [88]. From the assertive tones and decisive contents of this vast production, it would seem that the future of our culture is under control. On the other hand, the current status of our historic centers and outskirts, in other words of our cities, tell us different stories that can be summarized in one judgment citing Zygmunt Bauman “. . .cities have become the landfill for the problems caused by globalisation” [89]. In the meantime, in the debate on the issues related to the digital future, the scientific community – who is in charge of the conservation and more generally the safeguarding of cultural heritage and its value – appears to be still too careless, when not absent. This community seems to be rather close in a sort of deaf disciplinary fence.

1490

50.6

P. Salonia

Conclusion or Where Will Our Memory Be?

Given a globally problematic present, an indefinable future perspective – fideistically based on the benefits and endless effects of technology of the unknown outcomes – is lying before us. If we want to observe the reality with a secular eye, at present this technology is already proving to generate that conformism of decisions and passivity of behavior [90] which diverts otherwise principles, values, and feelings. To avoid possible misunderstandings, this chapter was not the place to sustain integralist and old-fashioned positions which deny the immeasurable advantages that technologies have determined in all sectors of our life, from health to scientific research, from work to free time. Technologies have undoubtedly contributed to a general improvement of the quality of life (even if according to latitudes). This chapter aimed to objectively discuss the real transformations caused by technology in societies and today (and tomorrow) citizens, starting from the capillary introduction of the digital in all aspects of people’s everyday life. Every age and social class is involved in this process, thanks to exogenous decisional dynamics which most of the time are merely financial. Moreover, the overdose of globalized, deafening information does not certainly increase our cognitive capacities. In the end, that information relegates us in individual solitudes, where we are constantly worried about being disconnected, and of losing segments of this unstoppable flow impossible to internalize through processes of critical understanding. More specifically, in this article new interrogatives have been posed, looking at how such scenario reflects on topics of meaning and values which cultural heritage and historic centers will be able to continue expressing in the development of tomorrow cities and societies. In one word, it has been considered necessary and important to wonder what the future of memory can (still) be, in the firm hope that the digital memory will not delete the historic human memory and, as a consequence, the civil consciousness. The aim is to better understand the society of information and its technologies to attempt to influence the future development of its declinations oriented to cultural assets in a smarter and more sustainable approach, with our ethical awareness and our moral behaviors. In the absence of instruments, in-depth analyses, and univocal answers, only some considerations can be suggested to stimulate the necessary discussions. Once restated that talking of conservation means to talk contextually of landscape conservation and that to refer to cultural heritage means to refer to the entire complex system of “city,” a possible address is the one indicated by Salvatore Settis when he sustains that: “. . .either our heritage in its entirety and the living fabric of the city and landscape become again a place of citizens’ self-awareness and a generator centre of energies for the polis, or their destiny is to perish.” Always citing Settis, “the crisis that we experience is an additional reason to reflect, with an eye looking at the past and one at the future, on the historic models of heritage conservation and their destiny. In order to return to the consolidated ancient model of conservation contextual of the landscape and heritage, the shine and outburst requested by circumstances, and our responsibility

50

Role of Digital Technologies, Robotics, and Augmented Realities

1491

towards future generations, it is necessary to submit ourselves to new questions, and tensions. In order to heritage conservation to still have a meaning and. . .a future in the city, it is absolutely necessary to know how to trigger two cultural processes. The first one. . .is the full historic-institutional awareness of the civil and social function of heritage in history. . .The second process. . . is the full reintegration of the culture of heritage safeguarding into the big cultural developments of our time” [91]. This invitation makes us reflect on the necessity to recompose that fracture between the human and the technological factors above discussed. To this end, Piero Dominici’s words offer help when he affirms that the future “is of whom will be able to recompose the fracture between human and technological, of whom will be able to define and rethink the complex relation between natural and artificial, of whom will combine (not separate) knowledge and skills, of whom will combine two cultures (humanistic and scientific) both in terms of education/training and of definition of profiles and professional skills” [90]. Furthermore, he writes: “to overcome the dichotomy between human formation and scientific formation is too important. . .because we are in an ipertechnological civilisation and because the implications of the iper-technological civilisation are social, political, cultural and concern everyone. . .the identities, the subjectivities, the life in its entirety. . .the change of paradigm is not a slogan. . .the extraordinary scientific discoveries of these years have forced us to review all categories, and the related operative definitions, starting from those of knowledge and life. . .to rethink the skills, the paths, the logics. . .the more interconnected and interdependent reality asks us to overcome these logics of separation. . ..a school and education of no quality are the prerequisites for a society destined to be always more deeply uneven. . .” [92]. Citing Max Weber, Piero Dominici concludes: “. . .the market, if left to manage itself, only knows a dignity of things and not of people. . .This is why it is necessary and urgent to do something about education and training” [92]. Therefore, a radical change of paradigm based on culture – and thus on education – and not on market is needed as an unquestionable generator of growth based on the not deferrable recovery of people’s own memory by citizens. People need to be able to regain possession of their identity as the only method to plan “new perspectives” for the twenty-first century, where words such as solidarity and inclusion, change and resilience, ecology and new consumptions, alike and diverse, assume substance and material. This will also be done by using the scenarios offered by innovative technologies, addressing them toward these objectives. In this perspective, historic cities and the entire heritage could be saved by the current progressive and unavoidable loss bringing them back to their role of driving force of social progress. Our historic cities need to become real laboratories of change, also in the name of beauty as an ethical and esthetical category, and as a practice because beauty is a fundamental need of our psyche, and it should be considered as an insuppressible necessity to live better, privilege of a common and shared serenity. Always Settis reminds us that “to think to the historic city means to think to human community, to the right of work and to the right of city. . .diversity and beauty are not a legacy of the past, but an extraordinary gift to the living present and an extraordinary gift to build upon and guarantee the future. . .” [78].

1492

P. Salonia

In order that cultural heritage and historic centers can continue playing their fundamental task for the growth of cities, and these can adapt to the changes brought by the technological progress in the twenty-first-century citizens’ consciences, it is now crucial and not more deferrable the development of an in-depth thinking. This really interdisciplinary thinking will have to involve, with the possibility of dialogue, urbanists, geographers, anthropologists, engineers, sociologists, architects, philosophers, with the common aim of designing new models relevant to the progressive non-stoppable evolution of the historic landscape. Undoubtedly, culture and technology are destined to create always much closer synergies – making the profound changes of paradigm necessary – which will be able to interpret an innovation more based on the merging of knowledge and increasingly pervasive ecosystem dynamics. In the twentieth century, regarding the innovative rousing aspects of that historic period, Richard Buckminster Fuller reminded everyone of us that “we are called to be architects of our future, not its victims” [77]. Here, an important moral imperative arises and which was absent in the deterministic perspectives: if we are in charge of shaping our future, we can only do it in the light of the personal responsibility, and therefore from an ethical point of view. To this end, with resolute awareness and willingness, we should be committed to developing “skills which will be useful not only at a local level, in the physical environment. . . but also at a global level. . .” so that “. . .we will be ready to challenge ourselves in the light of the big task that is awaiting us. . .the task of making the human community human” [89] – in harmony with the profound thinking left us from one of the greatest intellectuals of the twentieth century, Zygmunt Bauman. With the same ideals, Raffaello Sanzio wrote to Pope Leone X in 1519: “as the calamity of war generates the destruction and ruin of all disciplines, so peace and harmony generate the happiness of people and the laudable idleness. . .” [93].

References 1. Salonia P (2003) Strumenti informatici innovativi di ausilio alla conservazione del patrimonio storico-architettonico: problemi di organizzazione, diffusione e gestione dati. In: Rossi M, Salonia P (eds) Comunicazione multimediale per i beni culturali. Addison-Wesley, Milan, pp 25–52

This chapter organizes and expands the concepts expressed by the Author in previous printed papers related to some speeches presented during two Conferences at the University Federico II of Naples: - Salonia P, Patrimonio, Conoscenza, Tecnologie. Certezze o necessità di ripensamento critico? In Patrimonio culturale: tecniche innovative per il Progetto di Conservazione (edited by Rosa Anna Genovese). Proc. International Conference Patrimonio culturale, Città metropolitana, Paesaggio. Metodologie e tecniche innovative per la Conservazione ed il Restauro; Napoli, 2015; Ed. Giannini - Salonia P, Homo Digitalis e Patrimonio Storico: quale Futuro per la Memoria? In Patrimonio e città storiche come poli di integrazione sociale e culturale, Sostenibilità e Tecnologie Innovative. Esperienze internazionali di conservazione a confronto (edited by Rosa Anna Genovese). Proc. International Conference; Napoli, 2018; Ed. Giannini.

50

Role of Digital Technologies, Robotics, and Augmented Realities

1493

2. Drap P, Sgrenzaroli M, Canciani M, Cannata G, Seinturier J Laser scanning and close range photogrammetry: towards a single measuring tool dedicated to architecture and archaeology. In: Proc. 19th international symposium of the International Committee for Architectural Photogrammetry (Antalya, 30 September–04 October 2003) 3. Guarnieri A, Remondino F, Vettore A Photogrammetry and ground-based laser scanning: assessment of metric accuracy of the 3D model of Pozzoveggiani church. In: Proc. the Olympic spirit in surveying – International Federation of Surveyors working week (Athens, 22–27 May 2004) 4. Salonia P, Negri A, Valdarnini L, Scolastico S, Bellucci V Quick photogrammetric systems applied to documentation of cultural heritage: the example of Aosta roman city wall, in Cooperation to save the world’s Cultural Heritage. In: Proc. 20th international symposium of the International Committee for Architectural Photogrammetry (Turin, 26 September–1 October 2005) 5. Salonia P, Negri A, Valdarnini L, Scolastico S, Bellucci V Innovative quick photogrammetric systems for 3D cultural heritage documentation: the Appia Antica Project. In: Forte M (ed) Proc. the reconstruction of archaeological landscapes through digital technologies 2nd Italy-United States workshop, BAR international series 1379 (Rome, 3–5 November 2003) 6. Appolonia L, Moltedo L, Picco R, Salonia P Innovative systems for assisted analysis and diagnosis. In: Proc. 21st international symposium of the International Committee for Architectural Photogrammetry (Athens, 1–6 October 2007) 7. Salonia P, Scolastico S, Bellucci V, Marcolongo A, Letti Messina T 3D survey technologies for reconstruction, analysis and diagnosis in the conservation process of cultural heritage. In: Proc. 21st international symposium of the International Committee for Architectural Photogrammetry (Athens, 1–6 October 2007) 8. CENOBIUM. www.cenobium.isti.cnr.it/index.php 9. Salonia P, Scolastico S, Marcolongo A, Letti Messina T, Pozzi A Three focal photogrammetry application for multi-scale and multi-level cultural heritage survey, documentation and 3D reconstruction. In: Proc. 22nd international symposium of the International Committee for Architectural Photogrammetry (Kyoto, 11–15 October 2009) 10. Filas Regione Lazio. www.futouring.com/ 11. Salonia P, Letti Messina T, Marcolongo A, Appolonia L Photo scanner 3D survey for monitoring historical monuments. The case history of Porta Praetoria in Aosta. In: Proc. 23rd international symposium of the International Committee for Architectural Photogrammetry (Prague, 12–16 September 2011) 12. Studio di pre-fattibilità Pietrobon & Rossi Engineering s.r.l. : Tecnologie digitali per il restauro dei templi confuciani della provincia cinese dell’Hunan, OICE Progetto&Pubblico, vol February, p 11 (2010) 13. CNR: Relazioni degli Istituti – Patrimonio Culturale (2010) 14. ROVINA. www.rovina-project.eu/ 15. Stachniss C, Burgard W (2012) Particle filters for robot navigation. Found Trends Robot 3(4): 211–282. https://doi.org/10.1561/2300000013 16. Bogoslavskyi I, Vysotska O, Sarafin J, Grisetti G, Stachniss C Efficient trasversability analysis for mobile robots using the kinect sensor. In: Proc. European conference on mobile robots (Barcelona, 25–27 September 2013), pp 158–163 17. Kummerle R, Grisetti G, Strasdat H, Konoligie K, Burgard W A general framework for graph optimization. In: Proc. IEEE international conference on robotics & automation (Shanghai 9–13 May 2011), pp 3607–3613 18. Grisetti G, Kummerle R, Ni K Robust optimization of factor graphs by using condensed measurements. In: Proc. international conference on intelligent robots and systems (Vilamoura, 7–12 October 2012), pp 581–588 19. Basso F, Pretto A, Menegatti E Unsupervised intrinsic and extrinsic calibration of a cameradepth sensor couple. In: Proc. IEEE international conference on robotics & automation (Hong Kong, 31 May–7 June 2014), pp 6244–6249

1494

P. Salonia

20. Tedaldi A, Pretto A, Menegatti E A robust and easy to implement method for IMU calibration without external equipments. In: Proc. IEEE international conference on robotics & automation (Hong Kong, 31 May–7 June 2014), pp 3042–3049 21. Vergauwen M, Van Gool L (2006) Web-based 3D reconstruction service. Mach Vis Appl 17(6): 411–426 22. Moons T, Van Gool L, Vergauwen M (2009) 3D reconstruction from multiple images part 1: principles. Found Trends Comput Graph Vis 4(4):287–404 23. Hermans A, Floros G, Leibe B Dense 3D semantic mapping of indoor scenes from RGB- D images. In: Proc. IEEE international conference on robotics & automation, Hong Kong 31 May–7 June 2014, pp 2631–2638 24. Stachniss C, Burgard W. www.rovina-project.eu/del/del-1-1.pdf 25. Goody JR (ed) (2000) Il potere della tradizione scritta, Bollati Boringhieri. Feltrinelli, Milan 26. McLuahnl M (ed) (1967) Gli strumenti del comunicare. Il Saggiatore, Milan 27. Brusaporci S (2013) Modellazione e rappresentazione digitale per i beni architettonici. In: Brusaporci S (ed) Modelli complessi per il patrimonio architettonico-urbano. Gangemi, Rome, pp 18–24 28. Benjamin W (ed) (2000) L’opera d’arte nell’epoca della sua riproducibilità tecnica. Einaudi Ed., Torino. Benjamin W (1969) The work of art in the age of mechanical reproduction. In: Arendt H (ed) Illuminations. Schocken Books, York, pp 1–26 29. ICOMOS. www.icomos.org/charters/nara-e.pdf 30. Wang N (1999) Rethinking authenticity in tourism experience. Ann Tour Res 26(2):349–370 31. Greenwood DJ (1977) Culture by the pound: an anthropological perspective on tourism as cultural commoditization. In: Smith VL (ed) Hosts and guests. University of Pennsylvania Press, Philadelphia, pp 129–139 32. Silverman H (2015) Heritage and authenticity. In: Waterton E, Watson S (eds) Handbook of contemporary heritage research. Palgrave Macmillian, New York, pp 69–88 33. Dant T (1999) Material culture in the social world: values, activities, lifestyles. Open University Press, Buckingham 34. Malafouris L (2004) The cognitive basis of material engagement: where brain, body and culture conflate. In: DeMarrais E, Gosden C, Renfrew C (eds) Rethinking materiality: the engagement of mind with the material world. McDonald Institute for Archaeological Research, Cambridge, pp 53–62 35. Forte M (2014) Archaeology. New perspectives and challenges. The example of Catalhoyuk. Near Eastern Archaeol 2:1–29. https://doi.org/10.13140/2.1.3285.0568 36. Forte M (2008) La villa di Livia, un percorso di ricerca di archeologia virtuale. Erma di Bretschneider, Rome 37. Galeazzi F, Di Giuseppantonio Di Franco P, Dell’Unto N (2010) 3D cybermaps of Western Han mural tombs, cyber-archaeology. BAR international series 2177, pp 97–108 38. Kenderdine S, Forte M, Camporesi C (2012) Rhizome of Western Han: an omnispatial theatre for archaeology. In: Mingquan Z, Romanowska I, Wu Z, Xu P, Verhagen P (eds) Proc. revive the past. computer applications and quantitative methods in archaeology – 39th international conference. Pallas Publications, Amsterdam, pp 141–158 39. Kenderdine S, Shaw J, Kocsis A Dramaturgies of PLACE: evaluation, embodiment and performance in PLACE-Hampi. In: Proc. international conference on advances in computer entertainment technology (Athens, 29–31 October 2009), pp 249–256 40. Levy T, Petrovic V, Wypych T, Gidding A, Knabb K, Hernandez D, Smith NG, Schlulz JP, Savage SH, Kuester F, Ben-Yosef E, Buitenhuys C, Barrett CJ, Najjar M, DeFanti T (2010) On-site digital archaeology 3.0 and cyber-archaeology: into the future of the past – new developments, delivery and the creation of a data avalanche, cyber-archaeology. BAR international series 2177. Oxford: Archaeopress, pp 135–53 41. ICOMOS. www.icomos.org/charters/venice_e.pdf 42. Brandi C (1977) Teoria del restauro. Einaudi, Turin 43. Alberti Leon Battista OG (1966) L’architettura (De re aedificatoria) (ed: Portoghesi P). Il Polifilo, Milano

50

Role of Digital Technologies, Robotics, and Augmented Realities

1495

44. London Charter. http://www.londoncharter.org 45. Niccolucci F, Beacham R, Hermon S, Denard H (2010) Five years after: the London Charter revisited. In: Artusi A, Joly M, Lucet G, Pitzalis D, Ribes A (eds) the 11th international symposium on virtual reality, archaeology, and cultural heritage. Eurographics Association, pp 101–104 46. Wurman RS (2011) To dream the waking dream in new ways. In: Brockman J (ed) How is the internet changing the way you think. Atlantic Books, London, p 24 47. Europeana Collections. https://www.europeana.eu/portal/en 48. Arches. https://www.archesproject.org 49. Cyark. www.cyark.org 50. Rekrei. https://projectmosul.org 51. Herculaneum Conservation Project. www.herculaneum.org 52. ICOMOS. www.icomos.org/charters/charters.pdf https://www.icomos.org/en/charters-andtexts/179-articles-en-francais/ressources/charters-and-standards/387-principles-for-the-record ing-of-monuments-groups-of-buildings-and-sites-1996 53. Dacos M. https://tcp.hypotheses.org/318 54. Grande A. http://sevilleprinciples.com 55. CIPA. https://www.cipaheritagedocumentation.org/cipa_ottawa/, https://www.isprs-ann-photo gramm-remote-sens-spatial-inf-sci.net/IV-2-W2/ 56. Michael Culture Association. www.michael-culture.eu 57. EPOCH. http://epoch-net.org 58. 3D-ICONS. http://3dicons-project.eu 59. European Commission. https://cordis.europa.eu/project/rcn/92569/factsheet/en 60. DARIAH-EU. www.dariah.eu 61. Cultural Heritage: a challenge for Europe. http://jpi-ch.eu 62. SEARCH. www.ub.edu/searchproject/the-project/the-7th-framework-programme/ 63. European Commission. https://ec.europa.eu/programmes/horizon2020/en 64. E-RIHS. www.e-rihs.eu 65. European Commission. https://ec.europa.eu/programmes/horizon2020/en/h2020-section/socie tal-challenges 66. UNESCO. https://whc.unesco.org 67. ICOMOS. www.icomos.org/images/DOCUMENTS/Secretariat/2015/GA_2014_results/ GA2014_Symposium_FlorenceDeclaration_IT_final_20150318.pdf 68. ICOMOS. www.icomos.org/en 69. UNESCO. https://whc.unesco.org/archive/convention-en.pdf 70. ICAR. www.icar.beniculturali.it/biblio/pdf/Studi/franceschini.pdf 71. MIBACT. www.beniculturali.it/mibac/export/MiBAC/sito-MiBAC/Contenuti/Norme-e-Pareri/ Evidenza/visualizza_asset.html_1095508472.html 72. Twenge JT (2019) More time on technology, less happiness? Associations between digitalmedia use and psychological well-being. Curr Dir Psychol Sci 28(4):372–379 73. Turk V (2017) Understanding generation alpha. WIRED Consulting 74. Presa Diretta. www.presadiretta.rai.it/dl/portali/site/puntata/ContentItem-60e66850-8f0b-4f3b807a-9068b533638b.html 75. We are social. https://wearesocial.com/blog/2018/01/global-digital-report-2018 76. Galimberti U (2000) Psiche e techne. L’uomo nell’età della tecnica. Feltrinelli Ed., Milano 77. Ratti C, Claudel M (2016) The city of tomorrow: sensors, networks, hackers, and the future of urban life. Yale University Press, Yale 78. Settis S (2014) Se Venezia muore. Einaudi, Turin 79. Calvino I (1972) Le città invisibili. Einaudi, Turin 80. Turnbridge JE, Ashworth GJ (1996) Dissonant heritage: the management of the past as a resource in conflict. Wiley, Chichester 81. Council of Europe. https://rm.coe.int/1680083746 82. UNESCO. https://whc.unesco.org/uploads/activities/documents/activity-646-2.pdf 83. UNESCO. https://whc.unesco.org/en/hul/

1496

P. Salonia

84. UNESCO. www.unesco.org/new/en/culture/themes/culture-and-development/hangzhou-congress 85. ICOMOS. https://www.icomos.org/images/DOCUMENTS/Secretariat/2015/GA_2014_ results/GA2014_Symposium_FlorenceDeclaration_EN_final_20150318.pdf 86. United Nations. www.un.org/sustainabledevelopment/cities/ 87. European Commission. https://ec.europa.eu/regional_policy/sources/policy/themes/urbandevelopment/agenda/pact-of-amsterdam.pdf 88. United Nations. http://habitat3.org/wp-content/uploads/NUA-English.pdf 89. Bauman Z (2005) Fiducia e paura nella citta? Bruno Mondadori, Milan 90. Carta M (2010) Dalla Carta di Machu Picchu all’agenda per le città del XXI secolo. In: Lima AI (ed) Per un’architettura come ecologia umana. Studiosi a confronto. Jaca Book, Milan, pp 204–215 91. Settis S. www.ilgiornaledellarte.com/articoli/2012/10/114543.html 92. De Carli S. www.vita.it/it/interview/2017/06/09/nella-societa-ipercomplessa-la-strategia-esaltare-le-separazioni/119/ 93. Paolo di Teodoro F (2015) La lettera a Leone X di raffaello e Baldassar Castiglione: un nuovo manoscritto. Annali della Scuola Nomale Superiore di Pisa serie 5 7(1):119–168

Virtual Historic Centers: Digital Representation of Archaeological Heritage

51

Maurice Murphy, Stephen Fai, Lara Chow, Eimear Meegan, Simona Scandurra, Sara Pavia, Anthony Corns, and John Cahil

Contents 51.1 51.2

51.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data Collection and Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.2.1 Terrestrial Laser Scanning (TLS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.2.2 Photogrammetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.2.3 Other Survey Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Historic Building Information Modelling and GIS (HBIM and HGIS) . . . . . . . . . . 51.3.1 Parametric Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.2 Automatic HBIM Generation from Point Clouds . . . . . . . . . . . . . . . . . . . . . . . 51.3.3 Procedural Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.4 Semiautomatic Procedural Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.5 Historic BIM Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.6 LOD AND LOA Specifications for HBIM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1498 1500 1500 1503 1505 1505 1506 1506 1507 1508 1510 1510

M. Murphy (*) · E. Meegan Virtual Building Lab, Dublin, Ireland e-mail: [email protected] S. Fai · L. Chow Carleton University, Ottawa, Canada e-mail: [email protected]; [email protected] S. Scandurra Polytechnic of Milan, Milan, Italy e-mail: [email protected] S. Pavia Trinity College Dublin, Dublin, Ireland e-mail: [email protected] A. Corns Discovery Programme, Dublin, Ireland e-mail: [email protected] J. Cahil Office of Public Works, Dublin, Ireland e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_51

1497

1498

M. Murphy et al.

51.4

1511 1512

Case Study: Virtual Historic Parliament and Precinct District in Ottawa . . . . . . . . 51.4.1 Canada’s Parliament Hill National Historic Site . . . . . . . . . . . . . . . . . . . . . . . . 51.4.2 Level of Detail (LOD), Level of Information (LOI), and Level of Accuracy (LOA) – LODIA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.4.3 West Block BIM (2013) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.4.4 Centre Block (2015) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.4.5 Library of Parliament BIM (2017) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.4.6 Conclusion: HBIM Quality Assurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.5 Conclusion: A Design Framework for Digital Representation of Virtual Historic Centers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.5.1 Systems Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1513 1515 1517 1520 1521 1522 1524 1526

Abstract

Digitization and virtual representation of archaeology and architectural heritage potentially connects tangible and intangible cultural assets allowing for recording, conserving, and documenting cultural heritage. Process workflow, case study, and design framework are presented for virtual historic centers intended for archiving, storage, and dissemination of heritage knowledge and information. The process commences with remote sensing and data capturing technologies such as terrestrial and aerial laser scanning, Global Positioning System (GPS), and digital photogrammetry. The resultant survey data is enriched with new methods for 3D modelling of historic environments based on heritage geographic information systems (HGIS) and historic building information modelling (HBIM). The enhancement of 3D data with semantic attributes as intelligent virtual representation of historic environments allows multiple user scenarios ranging from engineering conservation to education and knowledge extraction in addition to object visualization. Open access computing systems for large data management and dissemination are now being considered; these are based on game engine platforms and Oracle and PostgreSQL spatial databases, which are used for managing large datasets. Keywords

Historic building information modelling · HBIM · Digital surveying · Historic structures · Building conservation · Architectural conservation · Laser scanning · SFM · Photogrammetry · Virtual cultural heritage

51.1

Introduction

Virtual historic centers are proposed as dynamic repositories and portals to digitally represent and assemble connected tangible and intangible cultural assets for both historic urban and rural centers. The intelligent digital representation of architectural heritage, archaeology, and objects allows for multiple user scenarios ranging from conservation to education and knowledge extraction in addition to

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1499

object visualization. The combination of digital recording, modelling, and data management systems enables the interaction with complex, interlinked threedimensional structures containing rich and diverse underlying data. End users can encompass architectural and engineering conservation, education, and research in addition to public engagement and cultural tourism. It is essential to incorporate within design frameworks international principles concerning authenticity, integrity, and philosophical approaches such as those promoted in ICOMOS Charters [1] (Fig. 51.1). Initially in this chapter, a state of the art for digital data capture and modelling for virtual representation for archaeology and architectural heritage is presented. One case study is then presented: The Virtual Historic Parliament and Precinct District in Ottawa. The case study presents an ongoing design framework based on data collection and modelling of historic sites and structures using historic BIM. In conclusion, a design framework is presented for systems architecture and workflows for digital representation of virtual historic centers in archaeology and architectural heritage for archiving and storage and dissemination based on game engine platforms.

Fig. 51.1 Virtual historic Dublin – digital representation of architectural heritage and archaeology [2]

1500

51.2

M. Murphy et al.

Data Collection and Preprocessing

The use of terrestrial laser scanning (TLS), photogrammetry, and traditional surveying techniques in addition to processing the captured data is outlined in this section. At this point, 20 years of research and development have been carried out by scientists and engineers for the application of laser scanning and photogrammetry for digitally recording architectural heritage and archaeology. To understand the current approaches for recording historic assets, it is necessary to comprehend its evolution.

51.2.1 Terrestrial Laser Scanning (TLS) Terrestrial laser scanning (TLS) captures multiple points using a laser to measure distances and angles from the scanner sensor to an object that is being scanned with millimeter to centimeter accuracies being possible (see Fig. 51.2). TLS operates on three different principles: time of flight, triangulation, and structured light. All three types of laser scanners produce a 3D point cloud of the object. A time-of-flight scanner uses a laser light probe to detect the surface of an object and determines the distance between the object surface and the scanner through measuring the time from omitting the signal and receiving it when it returns. Angular measurements are recorded on the vertical and horizontal axes for each signal, and the xyz coordinates

Fig. 51.2 Data capture, terrestrial laser scanner is a device that automatically measures the threedimensional coordinates of a given region of an object’s surface [2]

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1501

Fig. 51.3 For higher accuracy, the process of structured light scanning involves projecting a known light pattern onto the subject, Corinthian Capital Survey, Four Courts Dublin [2]

are calculated as a single point and collected a point cloud for the scanned object. Triangulation scanners calculate 3D coordinate measurements by triangulation of the spot or stripe of a laser beam on an object’s surface that is recorded by one or more CCD (charge-coupled device) cameras and the sensor. The process of structured light scanning involves projecting a known light pattern onto the subject. The light patterns will deform in relation to the surface of the object. These deformations in light will then be picked up by two cameras placed each side of the light projector. Most modern scanning systems are fitted with a CCD (charge-coupled device) digital camera, and the image data that is captured can be used to color the product of the laser scan survey data (the point cloud). The prime factor in selecting a scanning method is the required accuracy and the distance and size of the object. While time-of-flight method is accurate, it suits measurement and recording of larger objects and environments over large distances. Triangulation and structured light scanners are much more accurate, achieving sub-millimeter accuracies. For smaller and more detailed scans, structured light or triangulation is used. Structured light scanning is fast and very accurate, but this method requires a dark environment to produce the best results (see Fig. 51.3) [3]. Early works include Allen et al. [4], which includes the scanning of Beauvais Cathedral and dealt with issues of point cloud registration. Registration is the combination of two or more point clouds taken from different observation points or the referencing of the scanned object in a global or project coordinate system. This is achieved using tie and control points that are either features of the object (e.g., corners) or special targets (spheres, flat targets with high reflectivity), which are identifiable in the point cloud at the processing stage. Software for registering point clouds (see Fig. 51.4) usually facilitates registration by special targets or by overlapping point clouds or a combination of both [3]. In the case of large structures where the placing of targets is not always possible, known features on the object are used to fully transform and align the scans [3]. GPS can determine the coordinates of the laser scanner position allowing for the scans from

1502

M. Murphy et al.

The point cloud is the product of the laser scanner which is brought through a series of processing stages to develop a useable model

Registration is the combination of two or more point clouds taken from different observation points

Fig. 51.4 Registration – the combination of two or more point clouds [2]

each position to be brought into a common frame of reference in a global or project coordinate systems. Alternatively, registration by cloud matching is carried out by selecting a pair of partially overlapping scans and transforming one scan into alignment with the other using appropriate algorithms [4]. Early work [5] developed protocols for cleaning and removing erroneous data or artifacts such as reflections of the scan through objects before point clouds are registered. Once identified, reflections from objects in the background and in the space between the scanner and object, e.g., trees and other objects in the foreground, moving persons, or traffic and atmospheric effects such as dust, can be dealt with. Early research also resolved issues of accuracy of laser scanning concentrating on smaller cultural objects, which require very high scan resolution. This is best illustrated by Stanford University and the University of Washington [6] in the digitizing of the sculpting of the Renaissance artist Michelangelo. The triangulation scanner at a resolution of 1/4 mm captured detail of the geometry of the artist’s chisel marks. The problems of random errors and object occlusion in the laser scan survey can be greatly reduced by integrating other survey data. In addition, the level of detail can be enhanced for smaller features by introducing independent data collection based on digital photo modelling. Ground truth using other precise surveying instruments (e.g., total station) should be established and collected during the survey process to evaluate the accuracy of the laser and image survey data [3, 6–9]. The product of terrestrial laser scanning (TLS) is point clouds and in some cases registered images, and these require processing steps in order to generate products that can be used to create 3D CAD models. Data processing stages include segmenting point clouds and filtering out unwanted data. Automatic triangulation

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1503

Fig. 51.5 Mesh surface model from 3D point cloud – Leinster House Dublin [2]

of 3D points can also be carried out to create a mesh surface model from the 3D point cloud. This 3D mesh surface model (see Fig. 51.5) can then be used to generate orthographic images by combining the 3D surface model with 2D images. 3D mesh models can also be textured using referenced image data. 2D cut sections and 3D vectors can also be generated from the 3D point cloud or 3D surface model. Research over the last 20 years includes validation and identification of the most efficient and accurate processing pipeline and established good practice for determining accuracy and registration of scans and the potential for documentation of cultural heritage objects [10–15].

51.2.2 Photogrammetry Photogrammetric techniques use images taken at different viewpoints to record the 3D geometry of a building or object. Low-cost digital cameras, powerful computer processing, and the greater availability of commercial and open source photogrammetric software have changed the availability and use for photogrammetry. Digital photogrammetry can provide a point cloud, 3D solid model, and texturing based on high-quality imagery and color information. Modelling of large areas, buildings, and small objects can be produced by either aerial or close range (ground-based) photogrammetry. The main principles of digital photogrammetry are based on triangulation where lines of sight (rays) from two different camera locations are located on a common point on the object. The intersection of these rays determines the three-dimensional location of the point. Using this technique with two images is known as stereophotogrammetry, and where multiple images and camera positions are utilized, this is described as structure from motion. The term bundle adjustment is

1504

M. Murphy et al.

used to concurrently compute all the unknown parameters. The outputs from photogrammetric surveys are identical to products obtained from laser scanning and include orthographic images, point clouds, triangulated surface models, and also textured surface models [16]. Low-cost digital photogrammetry is based on hardware for data capture such as SLR digital cameras or in some cases off-the-shelf cameras. The acquired data is then processed, fused, and integrated using state-of-the-art photogrammetric and computer vision algorithms, which are readily available with software platforms Autodesk ReCap, Agisoft PhotoScan, Microsoft Photosynth, Micmac, Pix4D, etc. Geo-referencing image-based approaches are gaining momentum because they are much more cost-effective than laser-based or structured light scanning methods. No expensive equipment is required, but off-the-shelf digital cameras or even mobile phones and tablets can be appropriate to obtain 3D models from photo sequences. Single objects or small monuments can be acquired with a handheld camera [10] (Fig. 51.6). Postprocessing for close range digital photogrammetry includes stereo processing and multi-convergent processing (bundle adjustment). Common processing stages

Fig. 51.6 Digital data capture using photogrammetry structure from motion [2]

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1505

include selecting common feature points between images, calculating camera positions, orientations, distortions, and reconstructing 3D information by intersecting feature point locations [12, 13]. Improving accuracy requires human interaction in the preprocessing of image data, for example, the introduction of object measurement coordinates and scaling for real-world metrics.

51.2.3 Other Survey Techniques Traditional survey equipment such as total stations and GPS/GNSS equipment can provide very accurate measurements. These tools are much slower for data capture as each individual point is manually recorded. For large cultural heritage projects, millions of points are often required to accurately record a complex building or structure. Total station and GPS/GNSS methods can be employed to record accurate control points needed for registration of point clouds or ground truth data for confirming accuracy.

51.3

Historic Building Information Modelling and GIS (HBIM and HGIS)

Building information modelling (BIM) was developed for the design, build, and future management of new buildings and facilities. In BIM, the production of virtual models can automatically generate not only standard drawings and schedules but also provides for structural, economic, energy, and project management analytical data. BIM can automatically create cut-sections, elevations, details, and schedules in addition to orthographic projections and 3D models (wireframe or textured and animated). All these views are linked to the 3D model and automatically update in real time, so if a change is made in one view, all other views are also updated. This enables fast generation of detailed documentation required in the architecture engineering and construction industries [12–15]. Applying BIM for historic structures involves initially data capture of the geometry and texture for the structure using laser scanning or digital photogrammetry and converting the digital survey data to solid building information models (BIM). Two problems exist for researchers in generating historic BIM; the first is the absence of complex historic architectural elements in existing BIM libraries, and the second is a system for mapping the objects onto remotely sensed survey data. While these problems were initiated by the developments of historic BIM (HBIM) carried out in Carleton University and the Technological University of Dublin, research in HBIM is now extensive across the heritage research sector [17–23]. More recent work concentrates on automation in detecting objects and features in point clouds for improving the current slow process of converting unstructured point clouds into structured semantic BIM components [24–28] (Fig. 51.7).

1506

M. Murphy et al.

Fig. 51.7 The scan to historic building information modelling stages [2]

51.3.1 Parametric Modelling While BIM evolves from 3D CAD principles, its novelty is built on both featurebased modelling and parametric modelling. Feature based is an object-oriented approach where in addition to geometry, objects are intelligent and exist in a database or library with semantic attributes, i.e., a door objects or window objects, etc. In addition, the objects are coded to interact spatially; for example, if an object is revised in plan, it will also change to the new value in 3D and in other orthographic iterations. Objects also interact with each other; for example, a door or window placed in a wall will cut an opening in the wall. Parametric-based design is based on variable values such as dimensions, orientation and location, which can be revised for mapping and fitting within the point cloud or image survey. An object is created in relation to other objects in its class; for example, if the length or angular value for an element is revised, the other dependent elements will change to accommodate this. The rules that control the parameters can also assign operations to objects such as scale, extrude, revolve, and hide. These rules can be added, modified or removed. Parametric library objects (such as doors or windows) allow objects to be reused multiple times in a model or in many different models with varying parameters. This approach is very efficient for modelling elements that are repeated but may contain geometric variation between different instances [12–15, 17, 20] (Fig. 51.8).

51.3.2 Automatic HBIM Generation from Point Clouds Object recognition and feature extraction from an unstructured point cloud are being developed by researchers as an automatic process for the automatic generation of

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1507

Sample GDL Code Rafter wd colar = 10 ang = 30 !!!pitch angle th = .075 llrafters dp = .2 ! Icolars or ceiling joist centre = 1 ! Ijoist centres n = 10 111! number ofjoist anghip = 45 11 Material "Paint-02" Rotx 180 Roty ang Block .5*wd_colar/COS (ang), th, dp del 1

Fig. 51.8 Parametric roof model; the figure above illustrates the initial extruded rafter, which is combined with other elements to build the hipped roof; each element can be changed in shape position to create different iterations for many types of roofs [2]

structured BIM models from the point cloud data, [27–31]. By identifying automatically distinguished elements such as planes, surface models, openings, and 3D vectors from point cloud survey data, these objects are used as basis to plot the HBIM. The extracted objects are attached with semantic attributes and can also be converted into parametric building components, but this process requires human interaction. Historical architecture and archaeology contain complex shape, and geometry and decay require human intervention, which cannot be replicated by a machine [24].

51.3.3 Procedural Modelling Perhaps more promising for historic architecture and archaeology are the approaches of procedural modelling where generation of 3D objects and 2D shapes is based on computer algorithms and rules introduced by the user to generate automatically buildings and spaces from a grammar and vocabulary of shapes. Procedural modelling was originally developed for film and gaming industries and later adapted in GIS as CityEngine platform applied procedural modelling techniques for existing buildings [25, 32–35]. While the automatic generation of building is generated by rules that are initially defined by the user, unfortunately, the application of procedural modelling for converting laser scan surveys to historic BIM has limitations. Converting laser scan surveys to BIM requires human interaction to distinguish the

1508

M. Murphy et al.

variances and complexity of historic architecture and archaeology which the machine cannot identify without human supervision.

51.3.4 Semiautomatic Procedural Modelling In the work of Dore et al. [27], procedural modelling rules are developed in HBIM environments to accelerate the mapping process of parametric objects onto point clouds. Geometric and feature information is detected in point clouds and used as input for developing parameters for the intelligent library objects, which represent the architectural elements of buildings. In the case of historic structures, when decorative features are striped from buildings, the facades and roof structures are made up of much simpler geometric shapes. For example, a facade to a building consists of the wall structure with openings that are cut into the wall, and the openings contain doors or windows. In Fig. 51.9, a typical classical building facade stripped of ornament is illustrated; this consists of a panel with two large windows, one medium and one small. In Fig. 51.9, detail a illustrates the variables to establish geometric parameters, and detail b shows a partial GDL script for implementation for the variable geometry to build a wall panel. A single panel is constructed in Fig. 51.9c and finally a series of sub-routines, which brings the opening panels together to automate the panel with a

a.

Sample GDL Script // Single panel and window opes. END: !A= 1.5 !!! WINDOW WIDTH B = 3 !!! WINDOW HEIGHT C = 2 !!! DISTSNCE TO WINDOW BELOW D = 2 !!! DISTANCE TO WINDOW ABOVE E = 2 !!! DISTANCE TO WINDOW LEFT F = 2 !!! DISTANCE TO WINDOW RIGHT 100: cprism_ "Paint-02", "Paint-02", "Paint-02" 5+5, Th, 0,0, 15, 0, A*6, 15, A*4, A»6, 15, A*4,0, 15, 0,0, -1,b.

Fig. 51.9 Procedural modelling for a typical classical building facade [2]

c.

d.

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1509

Fig. 51.10 Shape arrangement for facade generation based on procedural rules [2]

set of openings, and these are brought together to form the panel facade. Further detail can facilitate the single panels to be adjusted for any opening size or distance between openings. In Fig. 51.10, the final shape vocabulary and arrangements for forming a historic structure are illustrated on the left side of the figure. The basic elements that make up the vocabulary of shapes include blocks forming walls, window openings, and window and door joinery as infill tiles; a panel containing a door opening. Additional elements which make up the door-case such as columns and pediments will automatically be revised to correspond to changes made to the doorcase. The block forms the walls of the structure, openings are cut and multiplied, and windows and doorcases are added to openings. On the right side of Fig. 51.10, the levels of detail are based on initially the block representing the walls followed by LOD 2 where openings are cut; building elements in LOD 3 are then added, for example, joinery, roof, classical details, etc. Shape elements to create detail for ashlar stone, moldings, and another wall geometry are added in LOD 4. A more detailed classical shape grammar and vocabulary is detailed in Fig. 51.11. The shape grammars can be applied to form numerous facade arrangements with variable values controlling dimensions and object position in 3D space [25]. Classical orders formulated the rules which govern the distribution and combination of parts and resulted in what is described as a grammar of ornament and composition. The elements (moldings, profiles, symbols, etc.) make up the architectural vocabulary. The rules of classical architecture can be described as a grammar. Shape grammars introduced the concept that buildings are based on different architectural styles and can be divided and represented by sets of basic shapes, which are a limited arrangement of shapes in three-dimensional Euclidean space. These shapes are governed by replacement rules whereby a shape can be changed or replaced by transformations and deformations. The shape commands combined with a library of primitives allow for all configurations of the classical orders in relation to uniform geometry. Nonuniform and organic shapes are developed through a series of

1510

Architectural Rules and Shape Grammars

M. Murphy et al.

Shape Arrangement

Fig. 51.11 Shape rules, grammars, and arrangement [2]

procedures using deformation and Boolean operations while attempting to maximize parametric content of the objects. These shapes are stored as individual parametric objects or combined to make larger objects in a library. When the parametric objects are used in the historic BIM platform, they can be transformed and deformed to match real-world requirements. In Fig. 51.11, the architectural rules are represented by architectural manuscripts on the left in the figure. In the center of the figure, shape grammars consist of vocabularies of primitive shapes that are combined using operations such as extrude, combine, replace, and deform to create library objects. The library objects are mapped onto the survey data according to the conditions for final shape arrangement [25].

51.3.5 Historic BIM Documentation Laser scanning and photogrammetry surveying systems for cultural heritage objects emphasize the collection and processing of data. As a result, the output is the accurate 3D model mainly suitable for visualization of a historic structure or artifact. On the other hand, BIM software platforms can automatically create cut sections, details, and schedules in addition to orthographic projections and 3D models (wire frame or textured and animated). The documentation for conservation of objects is achieved through producing 2D and 3D features, plans, sections, elevations, and 3D views. Conservation documentation can be automatically generated from completed HBIM models. Where conservation or restoration work is to be carried out on an object or structure, conventional orthographic or 3D survey engineering drawings are required (Figs. 51.12, 51.13, and 51.14).

51.3.6 LOD AND LOA Specifications for HBIM Existing standards for new buildings have not addressed the challenges for BIM in the context of architectural conservation or rehabilitation. Literature reviews

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1511

Fig. 51.12 HBIM documentation, ortho projects, and 3D details [2]

concerning HBIM have recognized the limited application of BIM for existing buildings. Since then, the application of and literature on BIM for historical buildings has been increasing rapidly. However, one of the fundamental problems, as outlined by several reviews, is the lack of agreed-upon HBIM standards and classifications. Further detail concerning level of detail is discussed and illustrated in the following case study: the HBIM of Ottawa Parliament Precinct [23].

51.4

Case Study: Virtual Historic Parliament and Precinct District in Ottawa

In 2012, Public Services and Procurement Canada (PSPC) and Carleton Immersive Media Studio (CIMS) began a research partnership to explore the application of digital technologies for architectural rehabilitation and heritage conservation. Our research has focused on the Parliament Buildings National Historic Site of Canada and has explored: methodologies for digitization (integrating photogrammetry and terrestrial laser scanning), building information modelling of historic structures (HBIM), digitally assisted fabrication (robotic milling and 3D printing), and digitally assisted storytelling (web, mobile, and virtual and augmented reality). In this section, the focus is placed on the evolution of our BIM practices on Parliament Hill stemming from the research initiative with PSPC, and more specifically the challenges in selecting the appropriate level of detail (LOD) and level of accuracy (LOA)/model tolerance are addressed. Establishing an appropriate LOD

1512

M. Murphy et al.

Fig. 51.13 HBIM documentation combining scan and HBIM [2]

and LOA is a crucial decision in defining the scope of a BIM project as it has a significant impact on model use, efficiency, and management. However, the complex geometry and deformations often found in existing and historical buildings make it difficult to adopt existing BIM standards that have been developed for new construction. Our study will revolve around three of the four heritage buildings situated on the Hill – West Block, Centre Block, and the Library of Parliament. A detailed analysis of the scope of the project, data management practices, and modelling methodology will demonstrate an evolution of modelling practices and workflows leading to best practices and lessons learned.

51.4.1 Canada’s Parliament Hill National Historic Site The Parliament Hill National Historic Site of Canada is comprised of the Centre Block, East Block, West Block, and the Library of Parliament and is Canada’s most recognized national monument. As both the political and symbolic locus of Canada’s parliamentary democracy, the site is in every sense a stage where Canada’s nationhood is played out for national and international audiences. Construction of the Parliament Buildings began in 1859, and in 1866, they were officially opened to the

Virtual Historic Centers: Digital Representation of Archaeological Heritage

ACCURACY

1513

Blue accurate match Red high levels error in match

51

The Point Cloud is compared with the HBIM for accuracy Fig. 51.14 Accuracy measurement documentation – comparing HBIM and point cloud [27]

public. In 1916, tragedy struck when the original Centre Block building was destroyed by fire. Reconstruction of Centre Block began immediately with the new design developed by architects John A. Pearson and Jean-Omer Marchand. The first sitting of Parliament in the new building occurred 4 years later, but it was not until 1927 that the 98-meter (320-foot) Peace Tower was completed. Today, the Parliament Hill National Historic Site is admired for its exemplary Gothic Revival style. Both the grounds and buildings are recognized for their heritage significance and have been designated as Federal Heritage Buildings (FBHRO). A comprehensive rehabilitation program for Parliament Hill commenced in 2002 – beginning with the Library of Parliament – with the intention of repairing the historic fabric, modernizing services, and addressing changes in the functional program. Following the library, the rehabilitation of West Block began in 2011 and was completed in late 2018. The Centre Block program of work is now underway. The East Block will see two phases of rehabilitation – the first beginning in 2017 and the second phase will be started soon.

51.4.2 Level of Detail (LOD), Level of Information (LOI), and Level of Accuracy (LOA) – LODIA During the development of the West Block BIM (2013) and the initial stages of the Centre Block BIM (2015), LOD and LOA standards or guidelines for modelling heritage buildings did not exist. In order to establish consistency in our own

1514

M. Murphy et al.

work, CIMS began the development of internal guidelines in 2015. The three-tier, five-category level of development system borrows from the architecture, engineering, construction, and operations (AECO) industry standards and guidelines and indexes level of detail (LOD), level of information (LOI), and level of accuracy (LOA) – LODIA – for each element type. CIMS LOD describes graphical and geometric representation in a scale from simple placeholder to detailed model and is based on existing standards for new buildings (Fig. 51.15). The selection of a specific level of detail is determined by available sensor data and reference material and anticipated HBIM uses. The LOD breakdown is as follows: LOD 0 – the element may not be modelled and may be represented by a placeholder (e.g., point cloud, historical drawing). If modelled, the element is a generic form with nonspecific dimensions and geometry. LOD 1 – the model element shows the generic size and shape graphically but does not contain additional information such as material and detailing. For example, a window is represented as an outline only. It contains proper dimensions but does not show details. LOD 2 – the model element is represented graphically with primary materials shown. Connections and secondary materials are minimally represented. LOD 3 – the model element is accurately represented graphically. The material palette is shown, and connections are modelled – but fasteners are not. LOD 4 – the model element represents ALL graphical and geometric information, including fasteners and the size of individual members. This LOD is reserved for areas where comprehensive detail is required. In the context of the CIMS LODIA protocol, LOA reflects the level to which the deflection and deviation of the building element are modelled in the BIM. LOA is characterized as:

LOD 1

LOD 2

LOD 3

LOD 4

LOD 5

Levels of Detail = LOD Fig. 51.15 CIMS level of detail 0 to 4 of a window, CIMS LOA classification system – level of detail (LOD), level of information (LOI), and level of accuracy (LOA) – LODIA [23]

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1515

LOA 0 – no deflection/deviation is modelled. An average dimension is used for position and material thickness. LOA 1 – element deflection/deviation is modelled at corners and changes of materials. Deflection/deviation shown is in positioning and not in material thickness. LOA 2 – element deflection/deviation is modelled at a predetermined grid spacing (typically between 1000 mm and 3000 mm) and at corners or changes of material. Deflection/deviation is shown in positioning, not in material thickness. LOA 3 – element deflection/deviation is modelled at a predetermined grid spacing (typically between 300 mm and 1000 mm) and at corners or changes of material. Large deflections/deviations (typically 50 mm to 1000 mm) between grid spacing are added. Deflection/deviation is shown in positioning and in material thickness if the thickness shows a deflection/deviation greater than 25 mm. LOA 4 – the highest level of accuracy is accomplished through the creation of a mesh generated from point cloud data and contains all deflection and deformation.

51.4.3 West Block BIM (2013) In July 2013, an HBIM of the West Block building using geo-referenced point cloud data was started. At the time, the development of BIM for historical buildings of that scale was a novel idea – the potential uses, challenges, and best practices were unknown. Therefore, the intention was to investigate the potential value of digitization and HBIM technology for the documentation, rehabilitation, and long-term management of the Parliament Hill site and beyond. The primary data used to develop the West Block BIM was from geo-referenced point cloud data, augmented with a diverse set of secondary data including photographs, historical drawings, 2D CAD plans, 2D CAD elevations based on orthographic photographs, and total station surveys. The interior survey was completed using the Faro Focus 3D. The exterior was surveyed using the Leica C10, with supplementary data from the Faro. The point cloud data was supported by an extensive photographic record, field notes, and hand measures taken during site inspections. In instances where scanning was not possible, the BIM followed secondary sources such as the CAD files prepared by PSPC. To amplify the complexity of the project, the building was an active construction site, requiring several scanning campaigns to capture the building through the construction phases to completion. Due to construction, some areas were not accessible to scan – such as the stairwells, portions of the interior, and the North Facade – until after the construction was complete. In total, 1.5 terabytes of geo-referenced point cloud data were captured. To enhance workstation performance and decrease model file size, the model was also divided into several component models that, when linked together, created a federated model. The component models included roof, shell, interiors, slabs, and historic structure (Fig. 51.16).

1516

M. Murphy et al.

Fig. 51.16 Component models for West Block BIM as modelling progressed and exploded axonometric of the Centre Block BIM showing the linked models [23]

The first step in developing the West Block BIM was to integrate the geo-referenced point cloud data into AutoDesk Revit 2014 by linking the individual .pts. files. The point cloud data was then viewed in a series of 2D section, elevation, and plan views to trace the profiles of the building element geometry to the assigned LOD and LOA. Next, using Revit’s modelling tools, a solid 3D model element was created. For LOD 1, elements were modelled as a simple placeholder with a minimal level of detail. Secondary sources were relied on heavily, as most areas specified at LOD 1 were not captured in the initial set of point cloud data. For LOD 2, building element types were developed from point cloud data in Revit – such as a window – and used to produce a library of parametric families. These building elements were then parametrically adjusted to accommodate their location in the model. At LOD 3, building elements were meshed from sensor data to generate model-in-place elements that afforded very specific and accurate models of unique geometric characteristics. Regarding LOD and efficiency, modelling to the highest LOD capturing most deformations required using model in-place components such as meshes or segmented point clouds. Producing solid HBIM elements was a time-consuming

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1517

process but increased the functionality of the model. Modelling at a lower LOD produced a model with a lower LOA but improved model functionality. Another important observation regarding LOD and LOA is the quality of the survey data. During the modelling of West Block, the resolution of the point cloud data was not high enough to model at a high LOD without relying on secondary sources. Based on our experience, it is imperative that a detailed documentation strategy is undertaken prior to developing a BIM for heritage documentation, conservation, and management in order to determine and reconcile the LOD of both the survey and the model. Following the initial phase of modelling (2013–2015), the second phase of modelling began in the summer of 2015.

51.4.4 Centre Block (2015) The intention of creating an HBIM of the Centre Block was to facilitate aspects of an integrated project delivery (IPD) method for the Centre Block Program of Work. CIMS would develop the BIM and hand the model over to the AEC consultant team responsible for the rehabilitation work. In addition to capturing the existing conditions of the building, the model was developed in anticipation of specific model uses including, but not limited to, the generation of drawings, site analysis, design coordination, and design authoring. In order to meet these objectives, the appropriate LOD for each building element category required specification. Our initial proposal was to utilize a commonly accepted BIM specification classification system – the level of development specification developed in the United States by BIMFORUM. The system combined level of detail with level of information classifications into level of development. A simplified level of development system was established between PSPC and CIMS for the initial scope, assigning LOD to specific building element categories: LOD 300: exterior walls, roofs, foundations, structural elements (verified to point cloud). LOD 200: interior walls, stairways, slabs, structural elements (not verified to point cloud). LOD 300 was also assigned to spaces of significant heritage value such as the Senate Chamber, House of Commons chambers, Senate and House of Commons foyers, Rotunda, and Hall of Honour. The model tolerance for the Centre Block BIM was determined by comparing the deflection and deviation of the building element to point cloud data. It was determined that a tolerance of less than 25 mm of modelled elements to point cloud data was acceptable. Any deviations greater than 25 mm would be captured within the geometric representation of the building element. Additionally, when the deflection or deformation was determined to be worth noting – based off the defined LOA – it was recorded in a customized properties panel. Three categories were defined: lateral deviation, vertical deviation, and irregularity of geometry. Lateral and vertical

1518

M. Murphy et al.

deviation referred to changes in the profile, whereas irregularity of geometry referred to deviations such as missing or broken elements. The Centre Block BIM required the synthesis of large, diverse data sets. The primary data source was geo-referenced point cloud data from terrestrial laser scanning and photogrammetry. The data was captured by CIMS using a Leica C10 and P40 (exterior and large interior spaces) and a Faro Focus (small to mid-sized interior spaces). Significant heritage interiors including the Senate, Senate Foyer, House of Commons, House of Commons Foyer, Rotunda, Hall of Honour, and the exterior of the Peace Tower were also captured by HCS using photogrammetry. Over 2000 individual scan stations were required to capture the interior and exterior of Centre Block – resulting in over four terabytes of point cloud data. Secondary sources such as archival drawings, photographs, historical steel catalogues, and technical reports were referenced in cases where point cloud data was not available. For example, the structural steel that is normally hidden from view and cannot be captured by laser scanning or photogrammetry. To increase workstation performance while modelling, the .pts. files were imported into Autodesk ReCap and divided into .rcp files by areas per floor, for instance, third floor south-west, third floor south-east, third floor East Office Block, etc. This way, model users could turn on/off specific areas of point cloud through the Revit Visibility and View settings for a small, localized area instead of loading in a large data set. Due to the size of the physical building, to minimize model file size, multiple component models when linked together created a federated model. The component models included roof, shell, interiors, circulation and slabs, and structure (Fig. 51.16). Similar in methodology to the West Block BIM, the first step in modelling Centre Block was integrating the geo-referenced point cloud data into Revit by linking the individual .rcp files. The point cloud data was then viewed in a series of 2D section, elevation, and plan views to trace the profiles of the building element geometry to the assigned LOD and LOA. Next, using Revit’s modelling tools, a solid 3D model element was developed into parametric families. As modelling progressed, we realized that BIMFORUM specification was insufficient for developing BIMs of historical buildings such as Centre Block. The availability of information for in situ building elements varied significantly, creating the need to identify levels of geometric detail, non-graphical information, and accuracy. In an effort to clarify the terms of reference, CIMS proposed the use of the CIMS LODIA – resulting in a more nuanced system of classification. The LODIA of each component was based on an understanding of the available data from the source material, as well as the anticipated use for that data. For example, the structural steel beams within the floor slabs could be identified as LOD 2, LOI 2, and LOA 0. The LOD and LOI were high because of the available archival drawings and historical catalogues. However, since there was no point cloud data available – most of the steel was behind masonry walls and not visible to remote sensors – the LOA was 0. In comparison, the Senate Chamber exhibited a different LODIA. While all specialty rooms in the Centre Block BIM were initially targeted to be modelled at

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1519

Fig. 51.17 Column developed at three LOD for the coarse, medium, and fine Revit detail levels [23]

a LOD 3, the Senate Chamber increased in graphical detail to a LOD 4 and LOA 4, while LOI was reduced to 2. This was because the model geometry of the Senate Chamber was leveraged for a virtual reality project related to the rehabilitation and required a high LOD for visualization purposes. High-resolution laser scanning and photogrammetry were used to record the space. However, very little verified information about wall assembly or materiality was available. A low LOD was required for conceptual/schematic planning, making the existing model too detailed, while a high LOD was required for visualizations and heritage asset management. One approach that we explored was developing model elements to three LODs. Using the Revit coarse, medium, fine detail settings, users were able to view all three LOD within one model depending on required information or use (Fig. 51.19). As the second phase of modelling progressed, the LOD and scope transformed prompting another division of the models due to increased file size. Model elements were divided into additional linking files including basement, heritage, courtyards, interiors (floors 1–3), and interiors (floors 4–6) (Fig. 51.17). The interdisciplinary research taking place between members of industry and academia has proven to be tremendously beneficial for both groups. The exchange of innovative workflows from research and standard practices from industry has pushed the application of digital tools for heritage conservation. The initiative is also supporting the development of highly qualified personnel (HQP) preparing researchers for both the intellectual and technical demands of a leadership role in defining the role of digital tools for heritage conservation in Canada and beyond.

1520

M. Murphy et al.

51.4.5 Library of Parliament BIM (2017) Our work on Parliament Hill culminated in the fall of 2017, with the digitization and modelling of one of Canada’s most significant heritage assets – the Library of Parliament consisted in developing a BIM of the library in anticipation of the following BIM uses: visualization for communication, scheduling/phase planning, site utilization planning, digital and digitally assisted fabrication, heritage asset management, field management, and record model. As a tool for heritage asset management and visualization purposes, the model required the highest level of detail and accuracy. Following research from the Centre Block BIM, it was determined that a tolerance of less than 25 mm of modelled elements to point cloud data was suitable for the Library BIM. In a similar methodology, any deviations greater than 25 mm would be captured within the geometric representation of the building element, and any deviations worth noting would be documented within the properties of an individual model element. The digitization program for the library was limited in scope. Only the main reading room, stairwells, attic, a few typical offices, and exterior were surveyed to produce geo-referenced point cloud data. We also requested highly detailed meshes of individual heritage assets in the main reading room – such as the handcarved wood rosettes and the statue of Queen Victoria – from photogrammetry since they would be required for the planned visualization applications. The data was captured using a Leica P40 (main reading room) and a Faro Focus (typical offices, staircases, and attic). One of the challenges in digitizing the main reading room was minimizing occlusions due to the room’s complex multilevel, circular geometry. The survey of the main reading room required 97 high-resolution scans in order to meet the required LOD and LOA for the BIM – taking approximately 5 days to complete. The exterior of the library was captured by HCS using UAV photogrammetry. We also relied on scan data from the Centre Block digitization campaign since the library data was geo-referenced to the same survey network as Centre Block. 2D CAD record drawings from the recent rehabilitation project (2007–2011) were referenced in cases where point cloud data was not available. However, we found significant discrepancies in instances where both point cloud data and record drawings existed. To increase workstation performance while modelling, the .pts. files were imported into Autodesk ReCap and divided into .rcp files by areas per floor. An exception to this was the main reading room where the 97 .rcs files were imported into Revit as individual scan station locations. The geometry of the room and file size of each scan made it difficult to group the scans into an effective and manageable . rcp file. From our previous research, we were confident that through proper Revit work-set management, we could contain the whole Library of Parliament building at a high LOD within a single Revit file. This eliminated the inefficiency of switching between models in order to adjust model elements – especially at the join conditions where linked files are connected. Similar in methodology to both the Centre Block and West Block BIM, the first step in modelling Centre Block was integrating the geo-referenced point cloud data

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1521

into Revit by linking the individual .rcp and .rcs files. The point cloud data was then viewed in a series of 2D section, elevation, and plan views to trace the profiles of the geometry of the building elements to the appropriate LOD and LOA. Next, using Revit’s modelling tools, solid 3D model elements were developed into parametric families. Despite our experience in modelling complex existing conditions, the geometry and detail of the library – including curved windows, elaborate fixtures, and flying buttresses – proved to be an extraordinary challenge. Modelling double curved geometry, intricate details, and surface deformations to a high level of detail from point cloud data required a return to exploring the possibilities of model-inplace families in special situations. For example, the geometry of the domed ceiling of the main reading room was extremely irregular. Our workflow involved producing a conceptual mass family generated directly from point cloud, then applying a generic model ceiling family, and deleting the conceptual mass. Our earlier research on Parliament Hill allowed us to evaluate the trade-offs between LOD and BIM performance, communicate data sources effectively, and use our existing protocols for modelling workflows outlining step-by-step instructions for modelling building elements from point cloud data. However, the increased complexity of the building required the augmentation of our existing protocols and development of novel ones to capture the LOD and LOA required for the specific BIM uses. This resulted in the Library of Parliament Model being our most complex BIM project to date.

51.4.6 Conclusion: HBIM Quality Assurance The process for verifying a model created from point cloud data involved creating multiple sectional views along elements in Revit and measuring the deviations that appeared to be the greatest between the point cloud and the model element. This method was time-consuming – notably for large BIM projects – and it limited the verification of the model to specific section locations. In the summer of 2018, a plug-in for Revit – 3D Analysis – was developed at CIMS (in association with the Photogrammetry and Geometrics Group, INSA). The plug-in is a first step toward an automated visualization of the deviations between Revit wall elements and adjacent point cloud data in a 3D view. After minimal computation time, points are colorized according to the computed deviations, and a 3D color map is displayed. To help the user analyze the deviations, a window containing information about the repartition of deviations is also displayed. The plug-in is proving more efficient in relation to the previous LOA verification processes. Moreover, deviations are represented in a 3D view making the identification of potential modelling errors and deviations more visible. Although we have achieved some success in automating the point cloud to BIM process, it must be acknowledged that the manual process used for the development of the Parliament Hill BIM has resulted in highly detailed and accurate models. Further, not all information for modelling a historic building is born digital. By synthesizing sensor data with secondary sources such as historical drawings, we

1522

M. Murphy et al.

have been able to generate a comprehensive representation of the fabric of the building. For example, the structural steel in Centre Block that is hidden from view within walls and floor slabs is now visible in the model and can be understood contextually. These secondary sources – integrated into the properties of the model elements for all three models – also offer the beginning of a rich database of non-geometrical cultural information related to the construction of the building (e.g., steel catalogues). As we develop LODIA workflows that produce more efficient and geometrically rich models, we continue to augment existing and create new protocols. The implementation of these protocols in the lab ensures consistency across all modelled elements in terms of modelling methodology. The ongoing work on LODIA is not intended as an attempt to develop an industry standard but rather as a forum for discourse and consensus building with our partners in a rapidly evolving field of research. Our intention in this chapter is to demonstrate to our public and private partners and to academic colleagues the value added for all parties in applied, collaborative research (Fig. 51.18).

51.5

Conclusion: A Design Framework for Digital Representation of Virtual Historic Centers

In conclusion, a design framework is presented for systems architecture and workflows for digital representation of virtual historic centers in archaeology and architectural heritage for archiving and storage and dissemination based on game engine platforms. Game engine platforms allow a low-cost method of making intelligent models and linked data more easily accessible to users. It is the nature of interactive video game applications to be intuitive to the user quickly upon assuming controls. A packaged “game file” is designed to execute in a standalone fashion, requiring no additional proprietary software installed on the end-user’s computer system. Current mainstream industry packages include Unity3D and Unreal Engine, which allow for highly developed workflows and community support, but recent game engine software like Autodesk’s Stingray package holds promise for greater interoperability with BIM. With regard to educational applications, game engines can give public access to information that is usually restricted to specialists. The nature of video game engines is scalable and multiplatform and can potentially be viewed on a variety of systems with different performance capabilities, from tablets to sophisticated virtual reality workstations. In addition, augmented reality (AR) and immersive experiences using wearable technology enhance the VR experience whether for entertainment or education. The virtual worlds are constructed in 3D graphic modelling platforms before they are exported into game engine platforms and only contain geometry and texture and are therefore limited to applications for visualization. The enhancement of the 3D visualization model for immersive experience with user interaction is generated within the game engine platform. This enables end users to interact with the virtual building and to access the rich data related to the model without needing to install specialist BIM

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1523

Fig. 51.18 (a and b) Historic BIM Parliament Precinct – graphic overview of scan and HBIM models [23]

1524

M. Murphy et al.

Fig. 51.19 Irish Parliament – Leinster House, aerial scan and HBIM model imported into Unity Game Engine Platform showing morphology of the parliament and ancillary buildings [2]

software. Shape, geometry, and geo-location can be linked to enriched data in the model and are held externally. User queries can be linked to the locations of elements in the building, to shapes (such as design features), and to the semantics of the information and will be facilitated by data flow between the game engine, 3D HBIM component server, and data stores. The delivery options for the Virtual Historic City range from WEB-based and VR immersion to augmented reality [36] (Fig. 51.19).

51.5.1 Systems Architecture A design system safeguards lasting value and lessens the risk of digital obsolescence. Open repositories ensure that curated data not only survives but is shared with wider communities for their use and enhancement avoiding duplication of effort. System design starts with the capture of data followed by its classification and organization. The organized data is then enriched with semantic attributes from other sources and stored within a database or repository allowing access for various end user scenarios. The entire workflow is continuously updated and improved through a continuing conceptualization, planning, and evaluation process and managed to ensure quality and survival of data. The initial design framework for virtual historic centers is presented in Fig. 20. Stage 1 illustrates the input data ranging from historic to laser scan and other survey data. This data is then processed and enriched with knowledge and information

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1525

Fig. 51.20 Systems architecture for virtual historic centers

attributes (stage 2) but also can be used in raw state. There are a series of database servers; the first is the historic components as 3D HBIM, which maintains the libraries of intelligent objects that represent the elements of a building structure. BIM authoring platforms are mostly tailored for modern architecture, and their libraries of parametric architectural elements/objects are limited to basic components. To overcome this problem, a new design model is applied using architectural rules and shape grammars to code primitive and complex historic architectural objects. The architectural objects are mapped onto a geometric framework made up of point cloud, image data, and historic digital surveys. A server is dedicated to game engine assets, and the system can also be linked into external data bases. The model at stage 3 holds AEC information and is then enhanced with geo-location, except for the 3D HBIM components server (detail 2); this can be considered a standard process pipeline. While BIM platforms have the potential to create a virtual and intelligent representation of a building, its full exploitation and use are restricted to a narrow set of expert users with access to costly hardware, software, and skills. In the final stage, the semantically enriched model is transferred into a WEB-based game engine platform. This not only enables interaction with the virtual building but also allows users to access and query related information rich

1526

M. Murphy et al.

data contained in the model and externally. The user access and queries are linked to the geo-location of elements of the building and to geometry shape and attributed semantics facilitated by data flow between the game engine asset server (detail 4) and the 3D HBIM component server (detail 2) (Fig. 51.20). Acknowledgments The authors wish to thank the Parliamentary Precinct Branch, Public Services and Procurement Canada, for their ongoing support of our research. They acknowledge Heritage Conservation Services, Real Property Branch, and Public Services and Procurement Canada for their technical support. This project was funded in part by the Social Sciences and Humanities Research Council (SSHRC) of Canada through the New Paradigm New Tools for Heritage Conservation in Canada internship program [37].

References 1. Charters Adopted By the General Assembly of ICOMOS., https://www.icomos.org/en/chartersand-other-doctrinal-texts. Accessed 8 May 2019 2. Figures developed from authors project archive material, Discovery Programme. www. discoveryprogramme.ie 3. Murphy M, Corns A, Cahill J, Eliashvili K, Chenau A, Pybus C, Shaw R, Devlin G, Deevy A, Truong-Hong L (2017) Developing historic building formation modelling guidelines and procedures for architectural heritage in Ireland. Int Arch Photogramm Remote Sens Spat Inf Sci ISPRS Arch 42:539–546 4. Allen PK, Troccoli A, Smith B, Murray S, Stamos I, Leordeanu M (2003) New methods for digital modeling of historic sites. IEEE Comput Graph Appl 23(6):32–41 5. Boehler W, Heinz G Documentation, surveying, photogrammetry. In: XVII CIPA symposium, Olinda, Brazil, 3rd–6th October 1999 6. Levoy M, Pulli K, Curless B, Rusinkiewicz S, Koller D, Pereira L, Ginzton M, Anderson S, Davis J, Ginsberg G, Shade J, Fulk D (2000) The Digital Michelangelo project: 3D scanning of large statues. Siggraph. In: Proceedings of the 27th annual conference on computer graphics. ACM Press/Addison-Wesley, New Orleans 7. Bernardini F, Rushmeier H (2002) The 3D model acquisition pipeline. Comput Graphics Forum 21(2):149–172 8. Beraldin JA (2004) Integration of laser scanning and close-range photogrammetry – last decade and beyond. In: XXth International Society for Photogrammetry and Remote Sensing (ISPRS) Congress Istanbul, Turkey, July 12–23, pp 972–983 9. Barber D, Mills J (2007) 3D laser scanning for heritage: advice and guidance to users on laser scanning in archaeology and architecture. English Heritage 10. El-Hakim S, Beraldin J-A, Remondino F, Picard M, Cournoyer L, Baltsavias M (2008) Using terrestrial laser scanning and digital images for 3D modeling of the Erechtheion, Acropolis of Athens. In: Proceedings of the Digital Media and its Applications in Cultural Heritage (DMACH 2008). 3–6 November 2008. NRC 50721 11. Patias P, Grussenmeyer P, Hanke K (2008) Applications in cultural heritage documentation. In: Advances in Photogrammetry, Remote Sensing and Spatial Information Sciences: 2008 ISPRS Congress Book, CRC Press, pp 381–402 12. Dore C (2017) Procedural historic building information modelling (HBIM) for recording and documenting European classical architecture. Doctoral thesis, Dublin Institute of Technology. https://arrow.dit.ie/builtdoc/17/ 13. Dore C, Murphy M (2017) Current state of the art historic building information modelling. Int Arch Photogramm Remote Sens Spat Inf Sci 42:185–192 14. Murphy M (2012) Historic building information modelling (HBIM) [For recording and documenting classical architecture in Dublin 1700 to 1830]. Doctoral thesis, Civil, Structural

51

Virtual Historic Centers: Digital Representation of Archaeological Heritage

1527

& Envi. Engineering, Trinity College Dublin. https://www.researchgate.net/publication/ 265396963_Historic_Building_Information_Modelling_HBIM_For_Recording_and_ Documenting_Classical_Architecture_in_Dublin_1700_to_1830_PhD_thesis 15. Murphy M, McGovern E, Pavia S (2013) Historic building information modelling – adding intelligence to laser and image based surveys of European classical architecture. ISPRS J Photogramm Remote Sens 76:89–102 16. Remondino F, Del Pizzo S, Kersten T, Troisi S (2012) Low-cost and open-source solutions for automated image orientation – a critical overview. In: Proc. EuroMed 2012 conference, LNCS, vol 7616, pp 40–54 17. De Luca L (2012) Methods, formalisms and tools for the semantic-based surveying and representation of architectural heritage. Appl Geomat (1866–9298). https://doi.org/10.1007/ s12518-011-0076-7 18. Baik A, Alitany A, Boehm J, Robson S (2014) Jeddah historical building information modelling JHBIM – object library. ISPRS Ann Photogramm Remote Sens Spatial Inf Sci II-5:41–47. https://doi.org/10.5194/isprsannals-II-5-41-2014 19. Boeykens S (2011, July 6–8) Using 3D design software, BIM and game engines for architectural historical reconstruction. Paper presented at the CAAD Futures 2011, Liège, Belgium 20. Chevrier C, Charbonneau N, Grussenmeyer P, Perrin J-P (2010) Parametric documenting of built heritage: 3D virtual reconstruction of architectural details. Int J Archit Comput 08(02):131–145 21. Fai S, Rafeiro J (2014) Establishing an appropriate level of detail (LoD) for a building information model (BIM) – West Block, Parliament Hill, Ottawa, Canada. ISPRS Ann Photogramm Remote Sens Spatial Inf Sci II-5:123–130. https://doi.org/10.5194/isprsannals-II-5-123 22. Fai S, Sydor M (2013, Oct 28 2013–Nov 1 2013) Building Information Modelling and the documentation of architectural heritage: between the ‘typical’ and the ‘specific’. Paper presented at the Digital Heritage International Congress (DigitalHeritage), 2013 23. Chow L, Graham K, Grunt T, Gallant M, Rafeiro J, Fai S (2019) The evolution of modelling practices on Canada’s Parliament Hill: an analysis of three significant heritage building information models (HBIM). Int Arch Photogramm Remote Sens Spat Inf Sci XLII-2 (W11):419–426 24. Garagnani S (2013) Building Information Modeling and real world knowledge: a methodological approach to accurate semantic documentation for the built environment. Paper presented at the Digital Heritage International Congress (DigitalHeritage), 2013 25. Hichri N, Stefani C, Veron P, Hamon G, De Luca L (2014) Review of the “as-built” BIM approaches. Journal of Applied Geomatics. Springer Berlin/Heidelberg 26. Oreni D, Brumana R, Della Torre S, Banfi F, Barazzetti L, Previtali M (2014) Survey turned into HBIM: the restoration and the work involved concerning the Basilica di Collemaggio after the earthquake (L’Aquila). ISPRS Ann Photogramm Remote Sens Spatial Inf Sci II-5:267 27. Dore C, Murphy M Semi-automatic generation of as-built BIM façade geometry from laser and image data. ITcon 19:20–46. http://www.itcon.org/2014/2 28. Jung J, Hong S, Jeong S, Kim S, Cho H, Hong S, Heo J (2014) Productive modeling for development of as-built BIM of existing indoor structures. Autom Constr 42:68–77 29. Volk R, Stengel J, Schultmann F (2014) Building information modeling (BIM) for existing buildings – literature review and future needs. Autom Constr 38:109–127 30. Xiong X, Adan A, Akinci B, Huber D (2013) Automatic creation of semantically rich 3D building models from laser scanner data. Autom Constr 31:325–337 31. Zhang R, Zakhor A (2014) Automatic identification of window regions on indoor point clouds using LiDAR and cameras. Paper presented to Applications of Computer Vision (WACV), 2014 IEEE Winter Conference on, 24–26 March 2014 32. Hohmann B, Krispel U, Havemann S, Fellner D (2009) CityFit – high-quality urban reconstructions by fitting shape grammars to images and derived textured point clouds. Paper presented to 3DARCH 2009 – 3D Virtual Reconstruction and Visualization of Complex Architectures, Trento, Italy, 25–28 February 2009 33. Muller P, Wonka P, Haegler S, Ulmer A, Gool LV (2006) Procedural modeling of buildings. ACM Trans Graph 25:614–623

1528

M. Murphy et al.

34. Parish YIH, Müller P (2001) Procedural modeling of cities. Paper presented to 28th annual conference on computer graphics and interactive techniques, New York, 2001 35. Thaller W, Krispel U, Havemann S, Redi I, Redi A, Fellner DW 2011) Developing parametric building models – the gandis use case. Paper presented to 4th international workshop 3D-ARCH 2011, “3D Virtual Reconstruction and Visualisation of Complex Architectures”, Trento, Italy, 2–4 March 36. Edwards G, Li H, Wang B (2015) BIM based collaborative and interactive design process using computer game engine for general end-users. Vis Eng 3(1):4 37. https://cims.carleton.ca/#/home

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural Heritage Applications

52

Adriana Bandiera and J-Angelo Beraldin

Contents 52.1 52.2 52.3 52.4 52.5 52.6

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . About Game-Changing Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Fortuitous Encounter or Improbable Collaborators . . . . . . . . . . . . . . . . . . . . . . . . . . . . A New Media Is Born . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Starting a 3D Imaging Laboratory from the Ground Up . . . . . . . . . . . . . . . . . . . . . . . . . Managing Challenges by Understanding Uncertainty Sources . . . . . . . . . . . . . . . . . . 52.6.1 Technological Strand: To Discard Data or Not . . . . . . . . . . . . . . . . . . . . . . . . . 52.6.2 Technological Strand: Best Practice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.6.3 Technological Strand: The Operator Is Part of the Measurement Chain! . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.6.4 Technological Strand: Should the Operator Breathe? . . . . . . . . . . . . . . . . . . . 52.6.5 Societal Strand: When Museum Curators and Users Are Excluded . . . . 52.7 Two Decades of Challenging (Hopefully Innovative) Cultural Heritage Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.8 Latest Cultural Heritage Projects at the SIBA, University of Salento . . . . . . . . . . . 52.8.1 High-Resolution 3D Models of Degradable Artifacts and Virtual Restoration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.8.2 3D Modelling for New Museum Use and Teaching Methods . . . . . . . . . . 52.8.3 The Muro Leccese Archaeological Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.8.4 The Second Life of a Unique Zygophyseter Varolai Specimen . . . . . . . . 52.8.5 Underwater Archaeology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.8.6 External Collaborations: The Bronzes of Punta del Serrone . . . . . . . . . . . 52.8.7 Internships as Prescribed by the University of Salento . . . . . . . . . . . . . . . . . 52.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix I – Applications of 3D to Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1531 1532 1532 1534 1538 1539 1540 1542 1543 1545 1546 1546 1547 1548 1550 1552 1553 1554 1557 1557 1560 1561 1563

A. Bandiera (*) University of Salento, Lecce, Italy e-mail: [email protected] J.-A. Beraldin Ottawa, Canada © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_52

1529

1530

A. Bandiera and J.-A. Beraldin

Abstract

In this chapter, the authors strived to provide a concise collection of key information on research methods, collaboration outcomes, and the importance of generating awareness in 3D imaging and modelling over a 20-year-period dedicated work. The aim of this chapter is to foster an exchange of ideas and information suitable for cultural heritage applications across several disciplines. Elements covered in this chapter will include a particular noninvasive technique known as 3D imaging and modelling, which is accepted as a fundamental approach to study and preserve cultural heritage. In addition, this chapter seeks to encourage multidisciplinary approaches that can assist site managers to better deal with their site from conservation to public outreach, demonstrate the importance and feasibility of post-processing complex image data by specialized technologists in a 3D laboratory, and outline the benefits of a 3D laboratory based on heterogenous skills. Abbreviations

(LE) 3D 3D48 3DTV ASCII BCE BDS BP BR CCD CD-ROM CE CEDAD CH CPU DH DIY DVD ENG EU FESR GPU ICP ISO IT MAUS

Province of Lecce, Italy Three-dimensional 3D for all Three-dimensional television American Standard Code for Information Interchange Before Common Era Brindisi Airport, Italy Best practices Province of Brindisi, Italy Charge-coupled device Compact disc read-only memory Common Era Centro di Fisica Applicata, Datazione e Diagnostica – Center for Applied Physics, Dating and Diagnostics Cultural heritage Central processing unit Digital humanities Do it yourself Digital versatile disc Engineering European Union Fondo Europeo di Sviluppo Regionale Graphics processing unit Iterative closest point International Standards Organization Information technology Museo dell’Ambiente dell’Università del Salento – Museum of the Environment of the University of Salento

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

MIUR NRCC NSF PON R&D RH RP S&T SfM SIBA SLR TPP WSA

52.1

1531

Ministero dell’Istruzione, dell’Università e della Ricerca – Italian Ministry of Education, Universities and Research National Research Council Canada National Science Foundation, USA Programmi Operativi Nazionali Research and development Relative humidity Rapid prototyping Science and technology Structure from motion Servizio Informatico Bibliotecario di Ateneo Single-lens reflex Teatro Pubblico Pugliese World Summit Award

Introduction

This chapter discusses how the introduction of digital three-dimensional (3D) imaging based on computer processing and stereo display systems has transformed the classical library service in an Italian university into a source of information augmented by interactive resources covering “built heritage” [1]. We assess applications based on systems that create dense digital 3D representations of the outer shell of objects or sites [2]. This is often referred to as 2.5D instead of 3D, and the term “3D scanning” is used for the creation of 3D data [3], although sometimes, no mechanical scanning is involved. In this chapter, we use 3D even though volumetric 3D has often been used in cultural heritage applications. Some authors use the expression “3D imaging” to describe only the visualization of 3D data, e.g., 3DTV and stereo displays. In the following sections, the expression “digital three-dimensional (3D) imaging” acquires all its significance. We use the term “3D imaging” to simplify the text. Moreover, this chapter aims to describe an almost two-decade-long collaboration between a technology-oriented research center in Canada (National Research Council of Canada – NRCC hereafter) and a library service in Italy, i.e., SIBA University of Salento that pushed the limit of 3D laser scanning and image-based techniques in the survey of built heritage. The expression cultural heritage is used afterward to signify material-based cultural heritage and “built heritage.” Digital three-dimensional imaging as a new form of media was shown to be a fundamental tool in the documentation (preservation), study (information extraction), and visualization (entry into the virtual world) of museum and heritage applications. Given that the collaboration started in 2000, the two teams had to face some important technological and human challenges, but the results were, nonetheless, successful demonstrations of democratization of technology and just how heterogenous knowledge could coalesce into a

1532

A. Bandiera and J.-A. Beraldin

versatile 3D laboratory dedicated to museum and heritage applications of 3D imaging. Our projects also tell stories, and these stories have the advantage of having known dates, readily available publications created by the present authors, and access to properly functioning and rich media creation tools.

52.2

About Game-Changing Technologies

The drastic evolution of computer systems, sensors, and community involvement in 3D imaging in the last 20 years has created a profound break with past methods to document and study cultural heritage. This fact may diminish the impact of the work presented here as one may forget that if one wants to innovate, standing still for the next best technology to be invented is pointless. As a society, we must explore and seek new ways with whatever material and knowledge we have at our disposal at a given time. The now normal daily use of 3D imaging techniques in all spheres of society was fostered by people that got their hands dirty, spent long nights staring at a computer screen, and got their fingers burnt on soldering irons putting systems together. What’s more, all the countless hours spent preaching to skeptical audiences where albeit one individual saw the potential has ultimately paid off. It is quite easy to find 3D imaging scientists who worked in the early 1980s, who tell stories of movie executives not believing in 3D models, manufacturing companies directors explaining that a caliper gauge is sufficient for their dimensional metrology tasks, and cultural heritage managers convinced that the digital world is just a fad. Those 3D imaging scientists were in fact the innovators in the “technology adoption life cycle” [4]. These misconceptions are normal and always appear in the early stages of new technologies as described in the “technology adoption life cycle” in the sociological model available from different sources [5, 6]. It seems that there is a continuous re-starting from the tabula rasa without any kind of assimilation of the culture that gradually takes shape and without having in mind the reasons that shifted the interest in favor of their use (introduction of digital technologies) from traditional techniques (mostly manual and material based).

52.3

A Fortuitous Encounter or Improbable Collaborators

Let us now look at almost 20 years of cooperation between the University of Salento and NRCC in the field of 3D imaging applied to cultural heritage. This chapter should demonstrate that when it comes to new technologies linked to new media, there is a natural progression of know-how from scientists/engineers to people that have little connection to those scientists/engineers’ skill set but can benefit greatly with any progressive/cutting-edge technology. The flow of know-how is not strictly unidirectional. Those scientists/engineers (often the innovators) come out of the collaboration with a better understanding of their own technology and fresh new avenues of R&D that they can apply to other fields of endeavor. It is like a

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1533

technology transfer or knowledge infusion to the actual users of 3D imaging for cultural heritage, e.g., predominantly people involved in digital humanity activities. The different ideas and skills coalesce into an interconnected technology that is now known as 3D imaging and modelling “3D imaging” for short. Three-dimensional imaging has rapidly become an indispensable and versatile documentation and analysis tool. The SIBA at the University of Salento applied for EU funding back in 1997, despite not yet having a 3D laboratory. Their role at the university was a traditional one: computer library services, digital library, scientific electronic publishing, and all related facilities for the academic and student populations. With the new funding from the EU in 1999, they were able to expand the development of the university digital library beyond traditional e-resources (databases, e-books, e-journals, etc.) into the creation and sharing of rich multimedia products (3D reconstructions, geo-referenced models, digitalized documentation of valuable objects and materials) for the preservation and promotion of cultural heritage material, both from within university walls and externally to national/international users. Let us now transition from Italy to Canada. As early as 1981, a group at the NRCC had been designing, building, and testing several 3D laser imaging systems and processing algorithms dedicated to the high-resolution modelling of complex objects and environments. The main interest was in the application areas as diverse as manufacturing, space robotics, and anthropometry. Initial demonstrations of these technologies to companies and potential clients were immediately appreciated in automotive and aerospace manufacturing, the area of dimensional inspection, in particular. A number of technology transfers were initiated as early as 1984 and resulted in the creation of a number of Canadian companies that now provide systems and services for countless applications requiring 3D imaging, modelling, inspection, and visualization. Many scientists and engineers were trained at NRCC in the early days of 3D, and many of them have since been evolving the academic and private sectors. In the early 1980s, NRCC realized that heritage conservation and documentation would provide a major avenue for research [7]. Industrial work being the principal mandate at the NRCC, scientific publications about industrial R&D were not always permitted. Heritage applications, especially those aimed at analytical tasks, pose high requirements on the data quality and the processing and visualization of the results. As it turns out, heritage applications were conducive to scientific publications. By their nature, museum collections provide a wide range of sizes, complexities, and materials for 3D imaging. Over the years, the group at NRCC initiated several pilot projects to gain a better understanding of the numerous issues in using the technology for heritage documentation. Each pilot project allowed testing and improving different aspects of the application of 3D imaging to museum objects in the laboratory and at remote sites. Their projects were conducted in collaboration with Canadian and international partners from museums and heritage organizations. Such collaborations provided invaluable firsthand experience and feedback in assessing the usefulness and relevance of the technology as well as a real-world context to identify possible technological improvements and further research

1534

A. Bandiera and J.-A. Beraldin

avenues. This dual use of a technology proved to be a winning combination. NRCC could test and improve its technology for industrial use, and cultural heritage organizations would access new technologies at a symbolic cost. As the NRCC technology was becoming publicized throughout the 1980s in television news, spontaneous demonstrations at Canadian museums (CMC [8], CHIN [9], RBCM [10], CANARIE [11]), and Canadian Conservation Institute (CCI [12]), a number of European museums and cultural organizations took note and contacted NRCC. This led to the creation of several pilot projects, which included the Louvre in Paris, France, and the British Museum in London, England. In addition, in 1999, NRCC hosted a Canada-Italy Workshop on the Heritage Applications of 3D Digital Imaging. The following year, in October 2000, the Embassy of Canada in Rome sponsored a “Canada Days – Giornate Canadesi” in Florence, which included a workshop on “New Imaging Technology for Cultural Heritage.” During that visit to Italy, NRCC was invited to present a tutorial on 3D imaging at the University of Salento. That initial visit would lead to many successful collaborations. The NRCC and the University of Salento (formerly University of Lecce) agreed to carry out a pre-competitive and non-binding cooperative program in the field of digital three-dimensional (3D) imaging applied to cultural heritage. This freedom allowed both laboratories to concentrate cooperation toward the application of innovative 3D sensing, processing, and visualization technologies to museum objects, paintings, archaeological site features, architectural elements, and sculptures. By applying the latest non-contact 3D technologies and methodologies, a bridge between expectations of art historians and curators (mainly academics), connoisseurs and the general public, in addition to the people promoting these new 3D technologies was created. The collaboration allowed both teams to demonstrate that a digital record or “digital model” can provide an archival quality documentation as important as the other documents generated in a cultural heritage project. This digital model can then be used for a variety of research, conservation, archaeological, and architectural applications and can create accurate replicas for interactive museum displays and virtual 3D theater applications.

52.4

A New Media Is Born

Three-dimensional (3D) imaging is nowadays a well-established form of media that encompasses acquisition, modelling, processing, database functions, and visualization. The source of data is a non-contact measurement process that samples a surface into a dense point cloud containing x, y, z coordinates and color information (appearance with proper computations). This allows for the possibility to store, transmit, and generate a model of the surface, enabling the model from arbitrary viewpoints and under arbitrary lighting conditions to open up countless possibilities. This media model is very similar to the one followed with textual information. In both cases, they are based on the previous generation’s skill set, i.e., the analogue pipeline is gradually replaced by an all-digital pipeline, a prospect only available

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1535

toward the latter part of the twentieth century. One should not forget that interest in digital 3D imaging actually started in the 1960s with the advent of readily available lasers, mini-computers, and vector-type displays. Semi-compact commercial systems became available in the 1980s with pioneering work in a number of private and publicly funded research laboratories around the world. Considering that 3D imaging captures three-dimensional information of people and physical objects and sites, the quest for quality in terms of “accurate” form (shape), size, and reflectance information measurement requires an adjustment to the media model. The model needs to be augmented by a metrological foundation or framework. Additionally, more and more practitioners are starting to discuss the legal aspects of exploiting accurate “surrogate digital models” of people, real objects, and sites. Figure 52.1 summarizes the media model for 3D imaging. No attempt is made here to explain in detail what is 3D imaging and what are all the applications of the technology, with exception of some details that pertain to cultural heritage, which are conveyed in Appendix I. The reason is fairly simple: three-dimensional imaging, modelling, and printing have become main street/stream technologies. Many laser and structured light-based 3D imaging systems have appeared on the market. Time-of-flight systems occupy the two-meter to several kilometer range, while triangulation-based systems are popular in the one-centimeter to two-meter range. The sub-centimeter domain belongs to interferometric type

Fig. 52.1 Three-dimensional (3D) imaging media model. We cover systems that create dense digital 3D representations of the outer shell of objects or sites. The term 2.5D may be used instead of 3D. Furthermore, “3D scanning” is also used as opposed to 3D acquisition. The latter term can also encompass volumetric 3D in which case not much is changed in the media model

1536

A. Bandiera and J.-A. Beraldin

systems with some competition by triangulation through the lens systems like confocal and focus-based 3D imaging systems. These systems cater to all sorts of applications and budget requirements. Within the three main class of 3D imaging systems, there are also different subclasses to systems tailor-made for specific applications. The latest arrival in the 3D systems tool box is in fact very old and very new. It is a 2D image-based modelling technique also known as multi-view 3D reconstruction. It uses photogrammetry based on the latest high-performance computer algorithms coupled to powerful personal computers based on multi-core CPUs and GPUs. Automated interest points detection, structure from motion (SfM) algorithms applied to sets of unordered images and dense image matching techniques deliver surface models approaching the pixel size of the images [13]. This latest technique and free open source (commercial solutions do exist) have helped practitioners in the cultural heritage field accomplish high-quality results at a fraction of the cost of the previous generation technology. The belief in 3D for all (3D48) is now achievable! The community has progressed toward the end tail of the innovation adoption life cycle. In fact, research interests are now directed toward extracting information from massive data sets; that is, data analytics, machine learning, big data visualization, and data-driven simulation dominate conference topics. These R&D activities are obviously attracting investments of more commercial nature that may trickle down to the cultural heritage sector. The reader can refer to this special issue of Springer Handbook on “Handbook of Cultural Heritage Analysis” for a rich applicative context where the results are brought to fruition for academics, connoisseurs, and the general public. Multiple books that identify the different types and technologies used for 3D imaging explain the physical and practical limitations when measuring surfaces in 3D; practical knowledge of the processing pipeline based on both professional and public domain software tools and strategies for applying metric tools to accurately measure and model an object/site exists. The reader can consult a number of books on the acquisition of surface data from the nano-scale to the kilometer-scale and on the processing of the 3D data in the terabyte range [7, 14–20]. An unrestricted search of library databases and the World Wide Web can produce an exponential quantity of publications on the different topics linked to the 3D media model. The restricted number of books listed in the reference section represent a good start to limit the search of informative material tailored to one’s research or applicative interests. We recognize that indirect methods to generate 3D data do exist. For instance, Fig. 52.2 shows the use of the transparency of a homogenous material carved to create a piece of jewelry. The grayscale information of a two-dimensional image (Fig. 52.2a) is converted to three-dimensional coordinates (Fig. 52.2c). That image is then rendered using synthetic shading (Fig. 52.2b, d). Though the results look visually pleasant, the 3D metrology behind the processing pipeline is quite ambiguous. In fact, this simple example can spur a series of questions. One could be satisfied with a visually pleasant result created from an opportunistic method to generate 3D data! This is when the mind of a scientist becomes inquisitive:

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

52

1537

c)

a)

b)

d)

Fig. 52.2 Three-dimensional data from the translucency of a Cameo (carving of marine shells’ surface). Rendering produced with permission by Diluca 1929 Classics S.r.l., Italia, www.diluca1929.it (2002), Cameo artist Francesca: Romana Titta, www.cameoitaliano.it (2019)

• • • • • • • • • •

What surface of the object did we measure and represent? What is the spatial resolution? What is measurement accuracy or uncertainty? Are the measurements repeatable, reproducible, or traceable? Was any documentary standard or certificate available? Are the digital data perennial? Was the 3D representation achieved at zero cost? Why should we be interested? Industries, academics, society in general? Who needs 3D imaging anyway? Is there either a commercial or academic target?

Why ask questions anyway? Answering these questions is critical for industry (as-built documentation, quality assurance, etc.). For instance, though the 1980s saw very little business for 3D imaging apart from dedicated systems for special applications where return on investment was almost instant, since the last economic downturn in 2008, the big 3D imaging players have been experiencing double digit growth [21–23], and a number of market studies have appeared sizing the current and future market value for 3D imaging technology in general [24, 25]. Therefore, there is an authentic market and it is growing. For cultural heritage institution, these questions become a matter of trust, measurement traceability, and perennial 3D digital models. The last topic will surely have to be address soon in order to make certain that the digital archive lasts through time and space. Our joint inquisitive minds at NRCC and the University of Salento luckily allowed us to innovate, contribute new knowledge in 3D imaging, and stimulate industrial and cultural interest. Therefore, the efforts of the collaboration have always been linked to metric 3D (defy the physical limitations), economic sense (useful to the community as a whole), and user value (fit for purpose).

1538

52.5

A. Bandiera and J.-A. Beraldin

Starting a 3D Imaging Laboratory from the Ground Up

The previous sections give the readers a context for 3D imaging and the parties involved in this collaboration, but we have not answered the question on how to reunite the interests, knowledge base, and skills of two seemingly different and already established teams. On one hand, there is a team that is engaged along a technological path “engineering” and a second, along a more “digital humanities” path [26]. Both teams had an advantage; they knew the value of both strands of endeavor, and they understood the importance of knowledge creation and transfer through information technology research. In reality, they were already engaged on a common goal to create 3D for all (3D48). The two strands (one technological and the other societal) finally met at the cross roads of knowledge (see Fig. 52.3), and both teams knew that cultural heritage applications of 3D imaging would help create spill over markets. To be precise, they promoted interdisciplinary work between the digital humanities (DH) and engineering (ENG) by applying some teamwork fundamentals: responsibility as individuals and at the leadership level, courage through self-leadership and never settle for second best and stay away from the “normalization of deviance” [27, 28]. Always question the work and the results (which have uncertainties) and let everyone have a voice (do not become a passenger on the project). Successful projects tend to have an antagonistic mode of operation. The individual is as important as the team. From experience, good ideas originate from individuals with different backgrounds and skill sets. Rarely does a bright idea come out of a committee meeting. Nevertheless, without true team brainstorming activities, individual ideas may lead to disaster. Individuals become collaborative at the execution stage, once an idea/method/process/algorithm is selected. One needs to listen but on occasion also needs to stop listening and push an idea to the team. Creating change or innovation is a difficult business. Change for the sake of change can be damaging to the whole project. In the course of the collaboration, this seemingly antagonist relationship fostered many innovations in 3D imaging.

Fig. 52.3 Two strands were joined and blended together: one technological (ENG-Engineering) and the other societal (DH – digital humanities). Both teams had a common goal to create 3D for all (3D48) with an application to CH (cultural heritage)

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1539

Fig. 52.4 A computer-based classroom used by undergraduate students in library technologies became a rendering farm at weekends. An example of how to be pragmatic and make use of what is available

In regard to the process of questioning the work and the results, if one asks a metrologist what the word uncertainty means, one will get an answer about measurement uncertainty [29] (i.e., the bread and butter of a metrologist). The answer would be as follows: measurement uncertainty means doubt about the validity of the result of the measurement, and it is a quantifiable property of a measurement. Uncertainty sources include standard(s) [30], workpiece (material), instrument (hardware/software), people (method/team), and environment (ambient conditions). If 3D imaging is based on the measurements of coordinates (x, y, z), then understanding the source of measurement uncertainty becomes a critical factor in every project. Everyday language prefers a term like accuracy. This term is treated as a qualitative term in metrology. International standards define a quantitative term measurement uncertainty to quantify measurement accuracy. Expressions like “it is good enough” or “it looks good” were not acceptable in the different projects we lead. The best practice section will provide more details. These considerations are in a way simple but have shown to be very effective and are part of our “Practical guide to starting a 3D imaging laboratory from scratch.” The other important ingredients for success are financing, purchasing, a business model for the whole collaboration and for each individual projects, effective project management (strategic planning), adequate workstation technology, high speed connectivity, trans-oceanic instant communication, staffing/training, best practice (processes and protocols), training, IT support, quality assurance, publications (paper, web, CD-ROM, tutorials, oral presentations), and numerous meetings with stakeholders to name a few. For instance, we needed to be pragmatic on several occasions and make use of what was available! One of the computer-based classrooms became a rendering farm at weekends (see Fig. 52.4).

52.6

Managing Challenges by Understanding Uncertainty Sources

We now turn to some of the pitfalls faced by the team in the course of the collaboration. Some of those pitfalls were outside of the team’s reach; others required the team’s complete attention and collaboration.

1540

A. Bandiera and J.-A. Beraldin

52.6.1 Technological Strand: To Discard Data or Not One of the first collaborative projects, which began in 2001, aimed at developing an effective approach for high-resolution photo-realistic texture mapping onto a dense 3D model generated from range images in the several meter range. The approach was applied to the virtualization of a Byzantine Crypt where geometrically correct texture mapping is essential to render the environment realistically. A combination of commercially available software was used, and custom-made 3D modelling software had to be developed [31]. Known as the Crypt of Santa Cristina of Carpignano Salentino (Italy), this ninth-century Crypt measures about 16.5 m  10 m  2.5 m and contains many of the region’s most ancient frescoes that are signed and dated, e.g., Christ and the Annunciation by Theophylact, 959 CE. A laser range scanner (Mensi SOISIC™ - now commercially obsolete) provided clouds of 3D points for the underground Crypt. Texture information was acquired separately with a high-resolution digital SLR camera (CCD resolution of 3008  1960 pixels – see Fig. 52.5b). One of the characteristics of the wide baseline triangulation-based Mensi SOISIC scanner was that the image intensity of the laser beam imaged on the internal camera was noisy and quantized with very few bits (see Fig. 52.5a). That intensity image being in perfect registration with the 3D data could have been used to register the high-resolution digital SLR camera images. As an alternative, synthetic shaded images of the 3D data had to be used instead, which created some added difficulties. A robust method to verify the registration had to be devised to reject mismatches between 2D and 3D images. A custom software was created for that purpose. The other alternative would have

a)

b)

Fig. 52.5 Comparison between the raw laser intensity on the return signal in the Mensi SOISIC scanner (a) and the high-resolution color image captured by a digital SLR camera (b). A) The point cloud (in SOI format) was rendered with the software package CloudCompareV2.10.2 (http://www. cloudcompare.org/doc/wiki/index.php?title¼FILE_I/O)

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1541

been to use the color video camera mounted inside the scanner. That solution was rejected after some accuracy tests. Many laser scanners before the year 2000 either generated low-quality laser intensity information or discarded that intensity altogether. Again, throwing away information is like ignoring a team member’s opinion. The second difficulty the team encountered was the fact that the manufacturer sold their own software and therefore they operated in a closed environment. Fortunately, an ASCII file was always generated along with the 3D point cloud in the proprietary format. That ASCII file was manageable because of its fairly small size. This was due to the scanner’s relatively low acquisition data rate (100 points per second) and the fact that the scanner could not operate for more than 1 h without rebooting the controller. Therefore, a complete scan of the whole environment, which would have taken 5–6 h, was not possible. It is worth noting that the latest laser scanners with measurement uncertainty ten times lower than the Mensi SOISIC can acquire data at a rate of ten million points per second. The third difficulty originated from the fact that the 3D coordinates were not organized in a regular scan grid (scan continuity was not preserved). To complete the processing pipeline, a software interface was designed to interpolate the point clouds using multiple planes of projection. This idea was then successfully implemented commercially with a package known as Innovmetric Polyworks™ [32]. The alignment technique based on ICP (iterative closest point) could then proceed normally, and it was finalized by a global alignment using the same software package. The Mensi SOISIC™ scanner came equipped with red sphere mounted on magnets that are suited for industrial sites where ferromagnetic surfaces are available. The Byzantine Crypt was excavated from limestone, and therefore, alignment spheres could only be placed on the floor. Different scenarios of sphere-based alignment were simulated and tested. The results were not as accurate as those with the ICP-based data-driven method. In another project, a Byzantine church created an interesting situation that was only just recently resolved. In any project, where gigabyte of data is generated, one will always ask the question “to discard data or not” especially if a critical situation occurs. For instance, the Mensi SOISIC had a major malfunction during the scorching summer months of the Salento region in Italy. The laser focusing mechanism melted and could not be repaired because that line of laser scanners was dropped by the manufacturer. What was left was a few 3D images and a lot of 2D photographs acquired with a digital SLR camera. Those 2D images were acquired to create a 3D model using a software application built around a multiview dense stereo 3D reconstruction algorithm. Many software packages have appeared since 2010. They are based on the latest work in photogrammetry and computer vison algorithms. We compare here 2002 technology versus 2019 technology based on a multi-view dense stereo 3D reconstruction algorithm. With the advent of digital acquisition of data, a lot more data can be acquired. The trick is to preserve the data and have a database where the information can be extracted. Otherwise, the sheer amount of data becomes overwhelming and also unusable in the long run (Fig. 52.6).

1542

A. Bandiera and J.-A. Beraldin

Fig. 52.6 Comparison between 2002 (top left and right) 3D reconstruction based on photogrammetry only and 2019 (bottom left and right) multi-view dense stereo 3D reconstruction based on photogrammetry and computer vison techniques. Top left: color texture 3D model with the wire mesh superimposed on it (about 1000 polygons), top right: synthetic shading showing the geometry based on simple shapes, bottom left: color texture 3D model without the wire mesh (3D model contains eight million polygons), bottom right: synthetic shading showing the more complex geometry (the color blue represents missing data)

52.6.2 Technological Strand: Best Practice The main objective of best practices (BP) is to increase the chances of success of a given project. Those commissioning (usually the end user) a project and those that actually execute (service provider) the work need to agree on a BP document, which outlines predefined specifications. We described in some details a two-layer BP structure that was successfully used in the Grotta dei Cervi project [33]. The key question is linked to minimizing the impact of measurement uncertainties on the amount of information desirable (fit for purpose) for a given deliverable. This question brings both the service provider and end user to understand the level of spatial resolution necessary given a budget and time frame and, at the same time, what is achievable by the physics of the measurement process. Figure 52.7 shows a schematic diagram summarizing the proposed best practices as applied to a painted cave project. Due to the complexity in terms of logistics, equipment, and people, the project Grotta dei Cervi required a very detailed plan. The BP documents needed to be followed attentively. Other projects required simpler BP, but nonetheless, the same structure was followed. Here, we provide the reader with an example on selecting data quality and quality assessment. The example refers to selection of the spatial resolution (lateral) for a 3D model of a bronze statue. The images below show some details of the statue and 3D model. A first trial resolution was

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1543

Fig. 52.7 Equipment used for the “Punta del Serrone” project: left) a laser scanner: Minolta Vivid 900 with a certified length ball-bar, right) low-cost dolly cart on which the heavy Minolta scannertripod combination was mounted

selected using the default parameter on the 3D camera. Two close-ups of the 3D model of the beard area are compared, one at the first trial resolution and the second at the higher resolution. Point density as a function of minimal feature size was tuned to deliver the required spatial resolution (lateral), and at the same time the return laser light saturation level of the laser scanner internal camera was adjusted to avoid clipping the laser intensity profile. Bronze affects greatly the performance of a laser scanner if used in default mode, which is usually tuned for Lambertian surfaces (Fig. 52.8). It is also important to show that reusing quality assessment equipment like a “ball-bar” (certified length implemented through a ball-bar) may not be adequate with a laser scanner based on a different measurement principle than the one it is intended for. Translucent materials create large biases with time-of-flight systems as shown in Fig. 52.9. Reusing reference artifacts for different scanners though apparently cheaper is not recommended unless proper metrological testing is carried out.

52.6.3 Technological Strand: The Operator Is Part of the Measurement Chain! We alluded earlier to the fact that the operator is part of the uncertainty chain. Fig. 52.10 shows the 3D modelling of a metope from a temple in Sicily. Starting from the same data originating from a laser scanner and the same 3D modelling software, a novice operator produced a mesh with 40,000 polygons (see Fig. 52.10a). An expert operator did the work and produced a mesh with 20,000 polygons (see Fig. 52.10b). The expert user has a knowledge of 3D laser scanners theory. The laser scanner in question produced raw data that

1544

A. Bandiera and J.-A. Beraldin

Fig. 52.8 Example of spatial resolution selection for a 3D model of a bronze statue. Top left: photograph of the head area, top center: view of the 3D model of the head at the first trial resolution, top right: view of the 3D model of the head at a higher resolution, bottom left: close-up photograph of the beard area, bottom center: close-up of the 3D model of the beard area at the first trial resolution, bottom right: close-up of the 3D model of the beard area at the higher resolution

Fig. 52.9 Some certified length standards. Left: two ceramic spheres with certified diameters mounted on a nominal 200-mm certified ball-bar, center: polymer sphere used with the Mensi SOISIC laser scanner – sphere diameter not certified, right: certified ball-bar length with two polymer spheres symmetrically mounted at the ends of that bar. The reference length provided with the Mensi SOISIC triangulation-based laser scanner was measured using a time-of-flight laser scanner. Laser penetration in the polymer caused the time-of-flight scanner to yield biased measurements

contained erroneous points especially on step edges and texture discontinuities. Failing to clean the raw point clouds prior to the modelling phase generates superfluous triangles in a mesh.

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

a) 40 000 polygons

1545

b) 20 000 polygons

Fig. 52.10 3D modelling of a metope from a temple in Sicily using same laser scanner and 3D modelling software. (a) A novice operator did not clean the raw 3D files from step edge effects before continuing into the modelling process phase. (b) An expert operator did clean the raw data using knowledge from 3D laser scanners theory – less superfluous triangles are generated

a)

b)

c)

Fig. 52.11 Systematic errors on 3D images. (a) Setup where an operator uses a 3D laser scanner mounted on a tripod that is resting with one leg on a scaffolding and two on a stable surface. (b) Wave phenomenon on the final 3D model created by the motion of the 3D laser scanner with respect to the object. Each color represents a 3D image after alignment in a common reference system. (c) Use of a mirror mounted on a second tripod – for the hard-to-reach surfaces

52.6.4 Technological Strand: Should the Operator Breathe? With older technology in the early years of our research, one had to make sure to be physically and mentally fit before using a combination of scaffolding, tripod, laser scanner, and a mirror mounted on another tripod (see Fig. 52.11). In this figure, we see the setup used to scan a metope. The operator had to stop breathing for about 30 seconds in order to allow for the stabilization of the scaffolding and the measurement time for each single laser scan. In Fig. 52.11b, each color represents a 3D image after alignment in a common reference system. Without that stability, the alignment of multiple 3D images yielded a wave phenomenon on the final 3D model. The problem is compounded if a mirror is added for the hard-to-reach surfaces. The latest lightweight handheld 3D scanners allow for the object and the scanner to move freely. The 3D model is created on the fly and available in real time. A simple web search with the keywords “lightweight handheld 3D scanners” leads to a list of systems tailored to different applications and budgets.

1546

A. Bandiera and J.-A. Beraldin

52.6.5 Societal Strand: When Museum Curators and Users Are Excluded Nowadays, one can apply the latest non-contact acquisition, 3D visualization, and interactive technologies to different cultural heritage situations with ease. In the end, a project may fail for a lack of consultation or consensus with museum curators or connoisseurs or the general public. This situation occurred in a large multi-team demonstration project where interactive museum displays were created based on the pre-conceived idea that technology alone is the solution to cultural heritage presentations. In a way, this was a situation similar to “ill-posed” problems. A solution was generated without asking some basic questions. Figure 52.12 shows visually what happened to a museum when the real expectations were not understood and met. In this particular situation, due to budget cuts, museum curators had no access to office computers, and therefore, the computers in the interactive displays were put to better use in their office. The lack of training of museum staff on the new digital displays became a defining issue in the project.

52.7

Two Decades of Challenging (Hopefully Innovative) Cultural Heritage Projects

The decades after the year 2000 were filled with innovative cultural heritage projects. The SIBA team went from rudimentary software kits to industrial-grade tools. As the same time, the NRCC team tested ideas, algorithms, and systems in beta development phases and was able to validate technology and innovate with new tools. Every project developed into a sort of value proposition for the team members and the stakeholders. A project would be accepted based on its relevancy, mostly if it delivered specific quantifiable benefits and unique differentiators. Indeed, these ideas were not off the shelf. The discussions were always in terms of a value proposition that people outside the circle of engineers and researchers could understand.

Fig. 52.12 When museum curators and users are excluded, interactive museum displays are bound to fail in a museum setting

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1547

In the 3D modelling pipeline, special attention was put on accuracy. Calibrated test objects and verifications through comparison of results obtained with other technologies like other 3D scanners, photogrammetric techniques, and coordinate measuring machines to name a few are recommended. This methodology became critical for obtaining high-quality reconstruction of 3D models from 3D imagery and spotting deficiencies early in the modelling pipeline. The investigation also centered around the effect that materials used for sculptures and paintings have on the accuracy of active and passive 3D techniques. NRCC worked on defining and assessing the level of dimensional and photometric accuracy required by museum curators. Here are some of the technologies that were tested and/or developed: • Atelier3D [34] facilitates or fully automates many operations, including registration of multiple scans, image processing before texture mapping, accurate registration of texture imagery with the geometric model, and interactive visualization of the full-resolution 3D model created from point clouds. Atelier3D allows viewdependent, real-time visualization of multiresolution models and compatibility with commercial 3D packages. • Il Teatro Virtuale [35] is a special software application for interactive visualization developed using scene graph tools. This tool was not only an instrument to visualize 3D results in real-time, but it is a successful tool for public outreach. • Other tools and technologies developed and tested include specific material standards to verify 3D scanner characteristics, measurement and modelling of 3D biases, modification of laser scanners to accommodate on-site 2D-3D calibration and texture mapping, tools to convert archaic 3D scanner data files into data format adequate for the modelling pipeline, and integration of point clouds generated from active (mainly laser-based) and passive (multi-view 3D reconstruction from 2D images) data into a seamless pipeline (see Fig. 52.13). These innovative tools were applied to multiple cultural heritage projects. The results are available in publications and limited edition CD-ROMs. Figure 52.14 shows the cover of some of the CD-ROMs.

52.8

Latest Cultural Heritage Projects at the SIBA, University of Salento

The projects carried out during the first decade of the collaboration benefited from substantial European funding for scientific research, technological development, higher education, and a climate that allowed NRCC to co-finance many science and technology (S&T) projects in many fields. After the EU funding ended and the NRCC mandate changed direction, what remained at the SIBA is a well-equipped laboratory and a fairly young staff with excellent training. NRCC researchers continued to collaborate but in a reduced way and mainly on feedback pertaining to research licenses used at SIBA. As a consequence, the activities of the SIBA 3D Laboratory have therefore been oriented toward internal users mostly at the

1548

A. Bandiera and J.-A. Beraldin

Fig. 52.13 Example of an integration of point clouds generated from active (mainly laser-based) and passive (multi-view 3D reconstruction from 2D images) data into a seamless pipeline. Rendering of one image used in a video animation produced for the project Land-Lab Selinunte [36]

Fig. 52.14 Cover pages: results from some of the projects were published in CD-ROM and DVD format [37–42]

university, with a particular focus on support for the research and teaching activities of the departments and museums of the University of Salento. These collaborations are carried out with modest financial means and without any large economic return. However, they represent an extraordinary wealth in terms of variety, experience, and results. The first applications after this important change were dictated by the requirements for conservation and virtual restoration made by academics. These projects were done in concomitance with more public outreach activities and student training. We will discuss some of these projects in the following section.

52.8.1 High-Resolution 3D Models of Degradable Artifacts and Virtual Restoration Artifacts found in an archaeological excavation are sometimes made of perishable or fragile material like wood, iron, or leather. These present understandable

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1549

conservation problems for scholars. Various artifacts with these features have been found in the Byzantine village of Scorpo, in the Supersano in southern Italy. This site has been the object of excavations by archaeologists from the University of Salento since 1999. On one hand, in 2007, a number of wooden objects were discovered in a well, while on the other hand, in 2012, a hoard of ferrous objects was found near a drystone wall that may have enclosed the settlement. Many questions arose: How should these highly fragile and perishable artifacts be studied? In what way should they be preserved for the future and in what manner should they be displayed to the public? A process ranging from 3D digital acquisition and modelling, digital restoration, the production of physical copies using a rapid prototyping equipment (RP), to their display in a museum was applied in this context. Resin replica of an object, created from its digital 3D model, is increasingly used in museums in order to improve communication, especially with visually impaired visitors. These replicas of fragile or in poor condition artifacts are useful for preservation and fruition purposes [43]. A crucial aspect of 3D modelling is the ability to perform restoration hypotheses in order to restore artifacts, especially those found in a fragmentary state, in order to look as similar as possible to the originals. Digital restoration does not result in any alteration of the original as it is performed on the polygonal model in a computer memory. Many scenarios can be applied to the same 3D model, and these can be evaluated by scholars before the actual physical restoration. In the case of a wooden cup, polygonal models of individual fragments were created, and a 3D model in which two fragments were assembled virtually was used for exhibition (see Fig. 52.15). In some situations, a complete artifact cannot be reconstructed from a physical point of view. This is the case of the sickle and the handle of the iron-made cauldron found at Scorpo. The first had been folded before its deposition and oxidation occurred on the contact faces which fused them together. In the latter case, the handle had two remaining fragments that could not be rebuilt physically because the Fig. 52.15 Restoration hypotheses of a wooden cup; left: 3D models of two fragments of the same object, right: three views of the assembled fragments

1550

A. Bandiera and J.-A. Beraldin

oxidation had, in a way, welded the attachment of the handle to the loop that was attached to the cauldron’s rim. The restorations performed on the digital models have allowed the virtual restitution of the sickle and of the fragments of the handle. Fig. 52.16 shows what the artifacts might have looked like. This allows more efficient research by scholars on the actual function and shapes of the objects in question. Starting from the 3D models, physical copies were made of resin (colored and finished manually), which were inserted into the permanent museum exhibition of Supersano (see Fig. 52.17).

52.8.2 3D Modelling for New Museum Use and Teaching Methods A number of initiatives to use 3D modelling in order to increase public outreach were launched with the MUSA (Historical and Archaeological Museum) of the University of Salento [44]. Digital technologies were used to improve the quality of the learning tools and communicative potential of exhibits. In particular, the application of 3D technology not only simplifies the detailed study of ancient artifacts but also increases the readability of an object, especially if the presentation

Fig. 52.16 Digital restoration of a sickle and handle found in an ancient Byzantine village of Scorpo (Supersano, southern Italy)

Fig. 52.17 Production of physical copies using a rapid prototyping equipment; (a) copies made of resin, (b) the real artifact near its replica. Artifacts found in an ancient Byzantine village of Scorpo (Supersano, southern Italy)

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1551

in a glass enclosure limits the visitor’s view or surface details are not visible with the naked eye (see Figs. 52.18 and 52.19). From a 3D model, one can build physical replicas or surrogate to the original in either full size or as a scaled version (see Fig. 52.20). These are used for educational activities with adults or children or visitors with visual or hearing impairments that require full engagement through tactile experiences. It is possible, finally, to create

Fig. 52.18 Cup with a bas-relief decoration of Corinthian production (Brindisi, Roman imperial age, 8.45  13.7 cm). The cup has a battle scene between knights. The convex surface of the vase and its location in a cabinet – showing only its front view – makes it impossible to follow the narrative of the bas-relief decoration. 3D digital modelling allowed for the unrolling of the scene

Fig. 52.19 Late antique amphorae from Vaste, Fondo Giuliano (southern Italy), 18  11,5 cm. 3D imaging and modelling have improved the understanding of an inscription on the surface of the small amphora. The digital model freed the engraved text from the painted decoration grid that covers the belly of the container, thus providing a better understanding of the text

a)

b)

c)

Fig. 52.20 Museum application of virtual restoration and display. (a) Mold for Eucharistic Bread (Byzantine era). (b) The 3D model facilitates the viewing of an animal figure engraved on the back of the object, which is not visible in the display window at the museum. (c) Virtual restoration of the Mold for Eucharistic Bread

1552

A. Bandiera and J.-A. Beraldin

animated digital walkthrough experiences for mobile learning or more generally for remote discovery. The project allowed the museum to reach a wider general public as well as implement an application for mobile devices and a catalogue of online searchable collections.

52.8.3 The Muro Leccese Archaeological Project The Muro Leccese Archaeological Project brought together the SIBA and a young and dynamic team from the Department of Cultural Heritage of the University of Salento that is conducting archaeological investigations at the Messapian site of Muro Leccese in Apulia, Italy. The collaboration has allowed the partners to experiment 3D imaging techniques for a new museum. During a period of more than 15 years of research on the archaeological site, a sector of the Messapian city was brought to light. In order to bring some innovation in the field of archaeological research and to expand the capabilities of the local museum, a series of 3D models for the analysis of ancient finds and for the digital restoration were created. The results obtained on a small scale are particularly important and innovative for both partners. A 3D imaging station was constructed to model very small ceramic fragments. One fragment of historical value that is characterized by a relatively small size (3.5 cm  4 cm  0.4 cm), with a slight curvature and specular paintings, was modelled in 3D. Two well-established 3D data acquisition techniques available at the university’s 3D laboratory were investigated for the system. The first acquisition technique is based on a high-resolution laser scanner. The second technique uses a multi-view dense stereo approach based on polarized light and a focus stacking process. Focus stacking is a convenient way to increase depth of field of a 2D image. Two metrically correct 3D models similar to the real artifact that are both functional and simple to display were created [45]. The digital model presented us with the possibility to identify the correct inclination of the fragment in order to identify the ceramic form (see Fig. 52.21). The second method provided the teams with higher resolution and more accurate results. This fragment consists of a small piece of a vessel with part of the shoulder and rim. The exterior decoration of the rim consists of a horizontal black band, whereas the shoulder has a black-figured band bordered by a black line in the lower section. The subject in the scene is a deer looking backward. The animal has been identified with a silhouette drawing and with some anatomic details incised. The side closest to the handle is limited by a palmette with nine petals and an empty heart (see Fig. 52.22a). The model was oriented using Autodesk 3DS Max and two parallel planes aligned with the decorations. Two sections were drawn, one on the top portion and one at the bottom. From the revolution of those sections, a circle placed on the top of the virtual cup was obtained with an estimated maximum diameter of 130 mm (see Fig. 52.22). The lack of the foot makes it difficult to identify its class, even though it is possible to hypothesize it is probably a piece of a skyphos (deep vessel for drinking wine).

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

a)

1553

b)

Fig. 52.21 Modelling small ceramic fragments. (a) Extraction of the main incisions on the digital 3D model created with a laser scanner. (b) Extraction of the main incisions on the digital 3D model created with a multi-view dense stereo method based on polarized light

a)

b)

c)

Fig. 52.22 The Muro Leccese Archaeological Project. (a) Rendering showing two parallel planes used to align the artifact with respect to a virtual cup, (b) and (c) creation of a virtual cup using Autodesk 3DS Max

52.8.4 The Second Life of a Unique Zygophyseter Varolai Specimen Another initiative that has led to significant results from the point of view of academic dissemination is the one arising from the collaboration between the MAUS (Museum of the Environment of the University of Salento) and the 3D Laboratory of the SIBA [46]. The collaboration that started in 2012 aimed at the creation of 3D models and digital content of fossil skeletal finds. The work has so far involved a series of fragments belonging to four ribs, five vertebrae, the mandible cut into four pieces, and two teeth of a Zygophyseter varolai that lived about ten million years ago. This ancient cetacean measures between 7 m and 10 m in length, and it is halfway between an orca and a sperm whale. It is the only specimen in the world of this type of mammal, and it was found in 1990 in a quarry in Cavallino in southern Italy. The remains of this Zygophyseter varolai is currently exhibited at the MAUS. The bone fragments show plastic deformations caused by the pressure exerted by the accumulation of overlying layers of material deposited over millions of years. A 3D virtual restoration (rectification) was therefore carried out (see Fig. 52.23a). A digital 3D model was also made of the 1:10 scale plaster reconstruction [47] of the same Zygophyseter (see Fig. 52.23b). The 3D models of the fossil fragments were then virtually positioned within the digital 3D model of the plaster reconstruction

1554

A. Bandiera and J.-A. Beraldin

a)

b)

d)

f)

c)

e)

g)

Fig. 52.23 The second life of a rare Zygophyseter varolai specimen. (a) Example of a 3D virtual restoration (rectification) of a bone fragment, (b) a digital 3D model was made of the 1:10 scale plaster reconstruction of the same Zygophyseter, (c) the 3D models of the fossil fragments were then virtually positioned within the digital 3D model of the plaster reconstruction that was first brought back to the right dimensional scale, in order to reconstruct the cetacean skeleton, (d–g) two virtual environments have been recreated, one submarine and one terrestrial

that was first brought back to the right dimensional scale, in order to reconstruct the cetacean skeleton (see Fig. 52.23c). Two virtual environments have been recreated, one submarine and one terrestrial (see Fig. 52.23d–g). They were used to create a narrative path in 3D computer graphics that describes the stages of the mammal’s history since it was alive until its discovery.

52.8.5 Underwater Archaeology The collaboration with the Chair of Underwater Classical Archaeology of the University of Salento started in December 2012. The main goal is the creation of 3D models and the eventual digital restoration of archaeological wooden elements found during the underwater excavations conducted by the University of Salento [48]. The university has been conducting numerous underwater archaeological works along the Ionian and Adriatic coast of Salento. Many of these sites have been composed of organic material remains, mainly wooden artifacts. In the absence of appropriate conservation and restoration operations, wooden artifacts deteriorate naturally through desiccation and bacterial and fungal decay. Even after

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1555

consolidation treatments, they require continuous monitoring. The techniques of wood restoration are expensive and usually irreversible. Therefore, we considered it necessary to immediately record and document the size of the wooden artifacts at the time of their recovery using 3D digital imaging technology. The first pilot project focused on a large wooden fragment from the underwater site of Santa Sabina (BR) of axial carpentry from a wreck (see Fig. 52.24 left). Radiocarbon dating by CEDAD (Centre for Dating and Diagnostics) at the University of Salento estimated a chronological window of 510–350 BCE (with a probability of 65.4%). However, it was necessary to prepare a protocol to be applied to materials requiring immediate restoration. In order to proceed with the 3D laser scans using a portable Minolta ® Vivid 900, the PhotoScanwater was first drained and the wood was left resting on the perforated base of the washing tank. In the following days, first in the alignment phase and later in the modelling phase, the presence in the 3D model of what we call “noisy points” and in some areas the presence of “double overlapping surfaces” indicated that the artifact had suffered dimensional changes throughout the day, revealing a general shrinkage but also bulging in certain areas. To determine with certainty whether the dimensional changes were caused by the progressive desiccation of the artifact or by a mechanical cause, e.g., the handling or crushing of the wood under its own weight, we decided to carry out tests on a smaller sample (approx. 15 cm. in length) that was detached from the rest of the main artifact of interest. The creation of the 3D models of the sample was performed with a higher-resolution (compared to the previous scanner) 3D laser scanner from ShapeGrabber. Figure 52.24 right shows the scanner and artifact setup. The application of 3D technology to underwater wooden artifacts has proved to be very useful. This work has allowed us to determine the time line within which one can operate out of the water without the artifact suffering significant morphological changes and to establish a protocol for all waterlogged wooden artifacts requiring documentation and restoration. In a second pilot project, the collaborators explored more flexible techniques in particular multi-view image-based dense stereo methods.

Fig. 52.24 Pilot project focused on a large wooden fragment of axial carpentry from a wreck which is probably a section of the keelson; left: the photograph shows the laser scanner used and the fragment in question; right: test sample positioned on top of a control object and on a rotating stage that automated the 3D acquisition phase

1556

A. Bandiera and J.-A. Beraldin

Many recent publications present solutions for the automatic generation of textured dense 3D surface models from 2D images [49]. A commercial software package Agisoft ® PhotoScan was selected for this experiment, and a public domain software CloudCompare was used for scaling purposes. The experiment for this pilot project were concentrated on a wreck, discovered in 2009 by a local diving team. It was found in the waters of Porto Cesareo, southern Italy, and at only 2 m underwater. It occupies an area measuring approximately 8 m  3.5 m. It is subject to continuous seasonal silting that has allowed its preservation. Radiocarbon dating performed by CEDAD (Centre for Dating and Diagnostics) at the University of Salento estimated a chronological window of 770–1030 CE with a reliability of 95.4%. In 2014, a stratigraphic excavation was performed, bringing to light more details of the whole wreck. During the excavation phase, all the fine sediments were suctioned with a special water pump to make the hull with its ballast visible. This allowed us to perform a 3D survey of the shipwreck before and after the ballast removal and the complete hull cleaning. The 3D survey was done by taking photographs with a non-metric camera inside an Ikelite diving housing. A Nikon D50 digital SLR camera and an 18–55-mm zoom lens with no optical image stabilizer were used. The photographs were taken in parallel rows at eight different angles with different convergence angles. A total of 600 images with an image resolution of 1504  1000 pixels were taken over the wreck at a standoff of about 1.5 m and used for the modelling phase in PhotoScan. Some thumb tacks and plastic cards were used for markings. They were used for the photogrammetric software and to identify certain assembly parts of the shipwreck. Pushpins were also used to indicate where wooden nails were previously located (Fig. 52.25). Scale bars were available, but for photo-aesthetic reasons, only manual measurements of different dimensions were carried out. This was contrary to what was instilled in previous projects and demonstrates that miscommunication is a considerable risk, regardless of a project’s size. As covered in a preceding section, a team member should have insisted to use scale bars with high accuracy. Scale bars should always be part of a photographic campaign. Manual measurements were used to scale the 3D model using CloudCompare with an estimated error of about 1%. The use of multi-view 3D reconstruction technology based on 2D still images in an underwater excavation proved to be beneficial. It drastically decreases the time spent in water and the time needed to document the wreck.

Fig. 52.25 View of the different camera locations: (a) side view of the texture mapped 3D model, (b) close-up showing a clearer picture of the wooden elements. Some thumb tacks and plastic cards used for markings are visible on the photograph

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1557

52.8.6 External Collaborations: The Bronzes of Punta del Serrone In 1992, just north of Brindisi at the bottom of the sea near the location “Punta del Serrone,” two hundred bronze fragments of statues, produced between the fourth and second centuries BCE and depicting divinities, philosophers, personalities in power, members of important families of antiquity, were discovered. The ship came from Greece where the original sculptures embellished prestigious public or private context connected to the figure of the richest intellectual and patron of the second century CE, Atticus Herodes. These recycled bronze statue fragments, meant for fusion into raw metal, were transported on a ship that, after finding itself in difficulty, had to throw its cargo overboard. Discovered among the sculptures bound for recycling was the only bronze statue thus far documented of the sophist Polydeukion’s favorite pupil. His image is present in the villas adorned with masterpieces that Atticus Herodes possessed in Greece but was also present in the sanctuary of Delphi [50, 51]. As part of a series of initiatives implemented by the Puglia Region (Italy) for the enhancement and promotion of the Apulian biblio-museum heritage in order to create a regional network of Biblio-Museums poles, which includes the involvement of the Pugliese Public Theatre – Regional Consortium for the Arts and Culture, an exhibition was created. The Brindisi Airport (BDS), and museums in Lecce and Brindisi hosted in 2019 the exhibition “Nel mare dell’intimità - L’archeologia subacquea racconta il Salento,” or “In the sea of intimacy - Underwater archaeology recounts Salento”, and at museums in Lecce and Brindisi. The Teatro Pubblico Pugliese (TPP) has asked the 3D Laboratory of the SIBA to create 3D models of some of the bronze statues from Punta del Serrone that are currently exhibited at the Archaeological Museum “F. Ribezzo” of Brindisi. The three-dimensional models are shown in the form of a photomontage (Fig. 52.26). Three-dimensional acquisition of the bronze statues was performed with a laser scanner (Minolta Vivid 900) equipped with two lenses, one for close-ups and one for the larger sections of the statues. To have more fluidity in the movements, a colleague of the SIBA with expertise in do-it-yourself (DIY) hobby skills has realized a low-cost dolly cart on which the heavy scanner-tripod combination has been mounted. This allowed a considerable saving of time when moving between scans. Figure 52.27 shows two photographs of the equipment.

52.8.7 Internships as Prescribed by the University of Salento In recent years, the requests to carry out training internships by students and especially undergraduates who want to access 3D technologies for the preparation of their thesis have multiplied. Access to SIBA to the internal activities of the university has favored the attendance by students to laboratory activities. They are always eager to acquire new skills, especially in the field of laser scanning and the creation of virtual 3D models. Although most requests come from humanities

1558

A. Bandiera and J.-A. Beraldin

Fig. 52.26 3D models of all the bronze statues from Punta del Serrone (Italy)

Fig. 52.27 Equipment used for the “Punta del Serrone” project: left) a laser scanner: Minolta Vivid 900 with a certified length ball-bar, right) low-cost dolly cart on which the heavy Minolta scannertripod combination was mounted

students (archaeology, physical anthropology, art history, etc.), engineering students are also interested in monitoring prototype material deformations. Each year, a number of engineering students attend the laboratory research sessions (see Fig. 52.28). During the internships, students are trained in the use of hardware and software equipment and are guided during all phases of the creation of a digital 3D model. The

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1559

Fig. 52.28 A graduate student from the physical anthropology field is shown here using the equipment at the SIBA of the University of Salento

Fig. 52.29 Mart3Di events: students have to illustrate the techniques used and discuss any critical issues specific to each artifact and the approach chosen to overcome them

activities cover various aspects, from the acquisition using laser scanners, photogrammetric techniques, to three-dimensional modelling, up to the creation of a model or of a video animation in computer graphics viewable on a giant stereoscopic display. At the end of their internship, the students elaborate, in the context of the event “Mart3Di,” a way of showing their work to the general public. These students have to illustrate the techniques used and discuss any critical issues specific to each artifact and the approach chosen to overcome them. The event held every last Tuesday (Martedi in Italian) of the month takes place in the 3D virtual theater. The idea of these events with the public has matured together with the students themselves, from the awareness that the documentation produced, thanks to 3D technologies and multimedia fruition techniques, is a highly innovative tool for the transmission of knowledge and a teaching aid that helps to encourage young researchers to pursue their passion and career (Fig. 52.29).

1560

52.9

A. Bandiera and J.-A. Beraldin

Conclusion

All beneficial and productive collaborations must come to an end (bad ones, however, are more easily forgotten). Three-dimensional imaging is now a wellestablished technology and is pointing to a bright future. From rudimentary 3D data acquisition and visualization in the early 1980s to a new era of information extraction and interpretation. The laboratory [52] in Italy is continuing the work independently in cultural heritage. The research laboratory in Canada [53] has transformed over the last few years and has morphed into a computer vision and graphics team that conducts applied research in the design of new and innovative 3D imaging systems, data analytics, machine learning, big data visualization, and data-driven simulation. Though these lines may elicit sadness from those who enjoy fieldwork, we can be reassured; the challenges faced by both teams produced results that are testimonies to many successful demonstrations of technological democratization [54] and deterritorialization. The exchange of information and skills from two very different partners confirms that heterogenous knowledge can coalesce into the creation of a versatile 3D laboratory dedicated to museum and heritage applications of 3D imaging operated by a new class of workers: digital techno-humanists. The inquisitive mind at the individual and team level allowed both teams to innovate, advance knowledge, and stimulate interests for 3D imaging applications in other areas. The thrust of the collaboration has always been one linked to metric 3D (challenge physical limitations), to selecting projects that made economic sense (useful to the partners, the science and technology community, and, to some extent, the community as a whole), and to include the end user (fit for purpose). Participation in international science and technology (S&T) projects is vital for any nation in order to access S&T knowledge; it needs to thrive in the current knowledge-based economy. In a matter of a few years, the SIBA of the University of Salento was able to reach a global audience of collaborators and at the same time brought home high-quality knowledge and skills that impacted the region of Salento in Italy. Both SIBA of the University of Salento and NRCC accessed the world’s best S&T facilities, equipment, and talent. The many collaborations extended from Canada to Europe and Australia provided and still provide vital access to the knowledge produced by researchers in other nations, a necessity for local stakeholders. It opened doors to a new type of employment, technology, and inspirational opportunities that are required to remain competitive. To be successful, any collaboration must have a win-win strategy and a conflict resolution process for the parties involved. Before the start of this collaboration, NRCC already had a network of collaborators, both academic and industrial, with many industrial licenses [55], publications [56], and awards. The collaboration allowed NRCC to expand its network of collaborators and interact with hard-to-reach industrial partners. Though the core group is composed of five people, SIBA of the University of Salento capitalized quite well with the collaborative work by attracting financing, ever challenging

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1561

projects with local partners and accessing crucial talents otherwise unaware of these opportunities in Southern Italy. The group continues to raise awareness and understanding of 3D imaging and virtual reality technologies in their local community as well as at international conferences. Over the years, the small group at the SIBA lab has worked with numerous internal and external collaborators, supplied specialized training to more than a hundred of undergraduate and graduate students, and has received accolades from national and international bodies like the World Summit Award (WSA), USA-National Science Foundation (NSF) Science and Engineering Visualization Challenge, and Italian Ministry of Education, Universities and Research (MIUR). The SIBA group continues to take on challenges with new projects mainly with local partners and companies. Every year university professors and students seek the help of both cultural and industrial projects. Their hard work and outreach efforts over the years have paid off. After completing their specialized work term at SIBA, students demonstrate their work during a special day open to the general public. This activity is a regular occurrence on campus. Acknowledgments Many of the projects were realized within Initiative 18 of the “Piano Coordinato delle Università di Catania e Lecce,” co-financed by the European Union (FESR, PON Ricerca 2000–2006), special funding from CASPUR-Rome (inter-university computing consortium), Activity 4 of the project LandLab (Laboratorio multimediale di ricerca, formazione e comunicazione sui paesaggi archeologici), co-financed by the European Union (PON 2000–2006, Ricerca Scientifica, Sviluppo Tecnologico, Alta Formazione), and many towns of the Salento region that provided access to many CH sites. The contributions of NRCC staff were instrumental in the different 3D campaigns. It is through their great dedication and important contributions that many of the projects ran smoothly. Special thanks go to many collaborators in Italy, the United Kingdom, the EU, Australia, Japan, and the USA. The authors would like to express their appreciation and sincere gratitude to the Superintendence for Archaeological Heritage of Puglia for allowing the teams to access many of the sites. Special thanks and appreciation go to Alexandra Beraldin, lecturer at l’École EAC-Marche de l’art, culture, patrimoine, luxe – Paris for her help in reviewing and editing this chapter.

Appendix I – Applications of 3D to Cultural Heritage Museum objects, paintings, archaeological site features, architectural elements, and sculptures can be digitized to provide a high-resolution 3D digital record of the object or site. The digital record, or digital model, should provide archival quality documentation which can be used for a variety of research, conservation, archaeological, and architectural applications, for fabricating accurate replicas as well as for interactive museum displays and virtual 3D theater applications. There is a real opportunity to open collections and information to a wider public through video games, mobile apps, etc. Museums can engage people without physical walls and borders. Experience and knowledge of own and other cultures can be accessed from anywhere and any platform. Telling stories from different point of views will be possible. The digital world may become an

1562

A. Bandiera and J.-A. Beraldin

ultimate equalizer; there is no need to belong to a prestigious institution to achieve quality work. Here are a variety of museum and heritage applications: • Archival documentation: tool mark details. • Museum exhibition and display applications: interactively examine fine details on 3D digital models using a large screen monitor. • Research applications: magnified, accurately measure, examined under different lighting conditions and display of 3D models with or without color, e.g., brush stroke on paintings, roll-out photographs. • Art conservation applications: the data provides an accurate record of the shape, surface condition, and color of an object, which can be used to document and monitor changes at different points in time. • Archaeological recording, architectural and historic building, digitizing sculpture applications: on-site acquisition and modelling. • Replication applications: the object is not touched or damaged during scanning; scale replicas can be made, which are much closer or truer representations than those copied by hand; the data can be formatted by machine the replica directly to make a mold. • 3D virtualized reality applications: interactive 3D virtualized reality systems offer the potential for the digital repatriation [57] of models of artifacts, which have been removed to distant museums, back into the virtualized model of their original site. High-resolution archival documentation Comparative conservation examinations Remote interactive 3D display Replication of fragile objects Examination of details Ethnographic collections Monuments Sculptures Paintings

Enhancement of museum exhibitions Virtualized 3D tours of closed sites 3D digital repatriation and restoration Archaeological site recording Art history research Natural history specimens Archaeological sites Historic buildings

Research community: CH provide Variety Shape and size (5-mm sculptures to cities) Materials Classes: stone, marble, ceramic, metal, glass, bone, etc. Metal types: gold, silver, iron/rust, brass, copper, etc. Environments: Sub divo, caves, dust/mud/snow/rain, heat/cold, etc.

Specific requirements on the performance and use of the technology Size of data set with respect to web/archive Impact on measurement: Translucent material Shiny/dark Dusty, etc. Engage people without walls and borders: Experience and knowledge of own and other cultures Tell stories from different point of views Spill-over activities: video gaming

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1563

Research community: R&D topics. Analyze Archival quality documentation: Storage, indexing, searching, and distribution Access to collection Closed or hard-to-reach sites Collection in museum vault Research: Examine and render fragile artifacts (complex shapes) Measure, process, categorize Conservation: Monitor deterioration Act upon condition ASAP Virtual restoration

Visualize Replication: Substitute Sale of low-resolution replicas Copyrights Interactive 3D VR theaters Make the experience alive through time and space! Collaborative environments Blend real world with digital world Web-based virtual museums Increase awareness of a site or an artifact Extend the life span of many exhibits

References 1. https://www.icomos.org/en/ 2. Jähne B, Haußecker H, Geißler P (eds) (1999) Handbook of computer vision and applications. Academic 3. https://en.wikipedia.org/wiki/3D_scanning. Last accessed April 2019 4. https://en.wikipedia.org/wiki/Technology_adoption_life_cycle. Last accessed April 2019 5. Arnold D, Geser G EPOCH research agenda for the applications of ICT to cultural heritage, May 2008. http://public-repository.epoch-net.org/publications/RES_AGENDA/research_ agenda.pdf. Last accessed April 2019 6. Arnold D (2017) Co-researching as a driver for technological innovation: computing and cultural heritage. LIBER Q 26(4):325–337. https://doi.org/10.18352/lq.10164. Last accessed April 2019 7. Bandiera A, Beraldin J-A, Gaiani M (2011) Birth and use of 3D imaging, modeling and visualization digital techniques for architecture and cultural heritage applications: a short history, in Italian (Nascita e utilizzo delle tecniche digitali di 3D imaging, modellazione e visualizzazione per l’architettura e i beni culturali), IKHNOS, Annali di Analisi grafica e Storia della Rappresentazione, Università degli Studi di Catania – Dipartimento di Architettura. Siracusa, pp 81–170 8. Canadian Museum of Civilization, now the Canadian Museum of History since 2012 9. Canadian Heritage Information Network 10. Royal British Columbia Museum 11. Formely the Canadian Network for the Advancement of Research, Industry and Education 12. Canadian Conservation Institute 13. Rupnik E, Daakir M, Pierrot Deseilligny M (2017) MicMac – a free, open-source solution for photogrammetry. Open Geospat Data Softw Stand 2:14. https://doi.org/10.1186/s40965-0170027-2 14. Liu, Y., Pears, N., Rosin, P.L., Huber, P. (Eds.), (2020) 3D Imaging, Analysis and Applications, Springer International Publishing, Springer Nature Switzerland AG 15. Luhman T, Robson S, Kyle S, Harley I (2011) Close range photogrammetry. Principles, techniques and applications. Whittles Publishing 16. Leach R (ed) (2011) Optical measurement of surface topography, 1st edn. Springer

1564

A. Bandiera and J.-A. Beraldin

17. Fryer J, Mitchell H, Chandler JH (eds) (2007) Applications of 3D measurement from images. Whittles Publishing 18. Chris McGlone J (ed) (2013) Manual of photogrammetry, 6th edn. ASPRS 19. Förstner W, Wrobel BP (2016) Photogrammetric computer vision. Springer 20. Ikeuchi K, Miyazaki D (eds) (2008) Digitally archiving cultural objects. Springer 21. InnovMetric Software Inc. news. https://www.innovmetric.com/en/innovmetric-reports-recordsales-3400-polyworksr-licenses-fiscal-year-2017-content-0. Last accessed March 2019 22. Faro news 2018. https://metrology.news/faro-report-double-digit-growth/. Last accessed March 2019 23. Creaform news 2011. https://www.creaform3d.com/en/news/2012-financial-results-creaformright-target. Last accessed March 2019 24. 3D imaging market size & share|growth analysis report by 2022. https://www.alliedmarke tresearch.com/3D-imaging-market. Last accessed March 2019 25. 3D imaging market – global industry analysis, size, share, growth, trends and forecast 2015– 2021. https://www.transparencymarketresearch.com/3d-imaging-market.html. Last accessed March 2019 26. Back in the 1990, the term “digital humanities” was not widely in use. People talked about “humanities” 27. “Countdown to teamwork” by Astronaut Mike Mullane. https://www.eharvard.org/Team/ Default.asp. Last accessed March 2019 28. “Petroski on engineering: the normalization of deviance”, by Henry Petroski. https://www. designnews.com/content/petroski-on-engineering-normalization-deviance/129953917932853. Last accessed March 2019 29. ISO 14253-2:2011, Geometrical product specifications (GPS) – inspection by measurement of workpieces and measuring equipment – part 2: guidance for the estimation of uncertainty in GPS measurement, in calibration of measuring equipment and in product verification (2011) 30. Beraldin JA, Mackinnon D, Cournoyer L (2015) Metrological characterization of 3D imaging systems: progress report on standards developments. In: Proc. 17th international congress of metrology, 13003. Paris 31. Beraldin JA, Picard M, El-Hakim SF, Godin G, Valzano V, Bandiera A (2003) Virtualizing a byzantine crypt: challenge and impact. Proc SPIE 5013:148–159 32. www.polyworks.com. Last accessed March 2019 33. Beraldin JA, Picard M, Bandiera A, Valzano V, Negro F (2011) Best practices for the 3D documentation of the Grotta dei Cervi of Porto Badisco, Italy, Proc. SPIE 7864, threedimensional imaging, interaction, and measurement, 78640J (27 January 2011). https://doi. org/10.1117/12.871211 34. Godin G, Borgeat L, Beraldin JA, Blais F (2010) Issues in acquiring, processing and visualizing large and detailed 3D models. In: Proc. 44th Annual Conference on Information Sciences and Systems (CISS) 35. Paquet E, Peters S (2002) Collaborative virtual environments infrastructure for E-business. In: Proc. international conference on infrastructure for e-business, e-education, eScience and e-medicine on the internet – SSGRRw02, January 21–27, L’Aquila, Italy 36. http://siba3.unile.it/land_lab/prog.htm. Last accessed April 2019 37. “Carpiniana: rappresentazione virtuale della cripta di Santa Cristina in Carpignano Salentino” [CD-ROM], Lecce, Coordinamento SIBA, 2002. ISBN 8883050061 38. “Carpiniana: a virtualized Byzantine Crypt” [DVD video], Lecce, Coordinamento SIBA, 2002. ISBN 8883050053 39. “Le Metope di Selinunte ¼ The Metopes of Selinunte” [CD-ROM], Lecce, Coordinamento SIBA, 2006. ISBN 8883050398 40. “L’Ipogeo delle Cariatidi di Vaste ¼ The Hypogeum of the Caryatids at Vaste” [DVD-ROM], Lecce, Coordinamento SIBA Università del Salento, 2010. ISBN 9788883050794 41. “Il Signore della folgore: Lo Zeus di Ugento ¼ Lord of sky and thunder: The Zeus from Ugento” [DVD-ROM], Lecce, Coordinamento SIBA Università del Salento, 2010. ISBN 9788883050787

52

Versatile 3D Laboratory: Challenging Aspects of 3D Imaging for Cultural. . .

1565

42. Divini eroi: un cratere da Cavallino e le sue storie ¼ Divine heroes: a krater from Cavallino and his tales ¼ Θει__κoί ήρωες. Eνας κρατήρας απó τo Καβαλλίνo και oι ιστoρίες τoυ [DVD video], Lecce, Coordinamento SIBA Università del Salento, 2009. ISBN 9788883050695 43. Bandiera A, Arthur P, Leo Imperiale M, Frigione M, Montagna F, Maffezzoli A, Signore GM Replicating degradable artefacts. A project for analysis and exhibition of Early Medieval objects from the Byzantine village at Scorpo (Supersano, Italy). In: Proc. 2013 Digital Heritage International Congress, Marseille (FR) 28 Oct–1 Nov 2013, pp 161–167 44. Signore GM, Bandiera A (2016) 3D imaging and new ways of making museums interactive and enabling digital discovery and learning. Museologia Scientifica 10 Nuova Serie: 129–136 45. Bandiera A, Meo F, Cammalleri A, Bianco C, Beraldin J-A Comparison of two well-established 3D acquisition techniques on a small fragmental artefact of a few cubic centimeters. In: Proc. 3rd Imeko International Conference on Metrology for Archaeology and Cultural Heritage, Lecce, 23–25 Oct. 2017, pp 477–482 46. Bandiera A (2015) La seconda vita dello Zygophyseter varolai, Balene e Barocco, suppl. of Thalassia Salentina n. 37, pp 29–35 47. The 1:10 scale plaster reconstruction of a rare Zygophyseter varolai specimen was made by the curators and restoration experts of the MAUS 48. Bandiera A, Alfonso C, Auriemma R Active and Passive 3D Imaging technologies applied to waterlogged wooden artifacts from shipwrecks. In: Proc. of TC V, CIPA Underwater 3D Recording and Modeling, vol XL-5/W5, 16–17 April 2015, Piano di Sorrento, Italy, pp 15–23 49. Remondino F, Del Pizzo S, Kersten TP, Troisi S Low-cost and open-source solutions for automated image orientation – a critical overview. In: Ioannides M, Fritsch D, Leissner J, Davies R, Remondino F, Caffo R (eds) Progress in cultural heritage preservation. EuroMed 2012. Lecture notes in computer science, vol 7616. Springer, Berlin/Heidelberg 50. Mannino K (2010) Inquadramento storico-artistico. In: Marinazzo A (ed) I Bronzi di Punta del Serrone. Dal mare al Museo Provinciale di Brindisi. Bari, pp 99–130 51. Mannino K (2013) Un gruppo di Erode Attico fra i bronzi di Punta del Serrone (Brindisi), Capolavori dell’Archeologia. Recuperi, ritrovamenti, confronti, Catalogo della mostra (Roma, 20 maggio – 5 novembre 2013), a cura di M.G. Bernardini, L. Lolli Ghetti. Roma, pp 219–223 52. https://siba.unisalento.it/3ddb. Last accessed April 2019 53. https://www.nrc-cnrc.gc.ca/eng/solutions/advisory/computer_vision_graphics.html. Last accessed April 2019 54. https://en.wikipedia.org/wiki/Democratization_of_technology. Last accessed April 2019 55. https://www.nrc-cnrc.gc.ca/eng/solutions/collaborative/research/collaboration_index.html. Last accessed April 2019 56. https://nrc-publications.canada.ca/eng/home. Last accessed April 2019 57. https://en.wikipedia.org/wiki/Digital_repatriation. Last accessed April 2019

3D Metrology for Ancient Pottery Classification and Reconstruction

53

Luca Di Angelo, Paolo Di Stefano, Anna Eva Morabito, and Caterina Pane

Contents 53.1

3D Geometric Model Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.1.1 Point Cloud Acquisition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.1.2 Point Cloud Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.1.3 Point Cloud Registration and Decimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.1.4 Point Cloud Tessellation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.1.5 Geometric Model Editing and Exportation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2 Geometric Model Fragment Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.1 Axis Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.2 Profile Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.3 Feature Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.4 Secondary Feature Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.5 Dimensional Analysis of Fragments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.3 Geometric Model Processing of Whole-Shape Pottery . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.4 3D Information Systems for Archaeological Pottery Visualization and Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.5 3D Puzzling of Archaeological Fragments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.6 Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.7 Additive Manufacturing Technologies for Physical Reconstruction of Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1568 1569 1571 1574 1575 1576 1577 1578 1580 1580 1582 1583 1584 1585 1586 1588 1590 1590

L. Di Angelo · P. Di Stefano (*) · C. Pane Department of Industrial and Information Engineering and Economics, University of L’Aquila, L’Aquila, Italy e-mail: [email protected]; [email protected]; [email protected] A. E. Morabito Department of Innovation Engineering, University of Salento, Lecce, Italy e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_53

1567

1568

L. Di Angelo et al.

Abstract

Ceramics classification and reconstruction are fundamental for the knowledge of history, economy, and art of a site. The method traditionally used by archeologists for their investigation presents a series of significant limitations. The results depend on subjectivity, specialization, personal skills, and professional experience of the operator; hence, they are not reproducible and repeatable. Furthermore, since the method is time-consuming, it is used to analyze only indicative samples that have characteristic components. In order to overcome these limitations, in the last years, some automatic methods for studying ancient pottery’s findings are proposed in literature. All the most promising ones analyze a 3D discrete geometric model of ceramics. By analyzing the voluminous related literature, the hottest topics are 3D geometric model setup, virtual prototyping, geometric model fragment processing, geometric model processing of whole-shape pottery, 3D puzzling of archeological fragments, classification, and additive manufacturing technologies for physical reconstruction of ceramics. In order to help all the researchers involved in this field, this chapter aims to provide a comprehensive and critical analysis of the state of the art for the abovementioned topics. For this purpose, the present review is focused on the presentation of the pros and cons of the techniques used on these different issues. Keywords

3D archaeology · Computer methods in archaeology · Surface segmentation · Axis evaluation · Automatic feature recognition · Computer-based methods for sherd classification and reconstruction

53.1

3D Geometric Model Setup

According to the United Nations Educational, Scientific and Cultural Organization (UNESCO), International Centre for the Study of the Preservation and Restoration of Cultural Property (ICCROM), and International Council on Monuments and Sites (ICOMOS) (1994), the primary objective of cultural heritage (CH) is the preservation of the authenticity and integrity of finds. The traditional techniques for performing these preservation activities in CH have proven to be time-consuming and not repeatable or reproducible, as they are dependent on operators’ skills. For these reasons, in the last 10 years, the traditional techniques have been gradually replaced by digital ones, which have also led to the introduction of new revolutionary approaches to preservation activities. These approaches need a 3D virtual model of the archeological find. The reconstruction of a 3D digital model of ceramic fragment or whole ceramic specifically requires different steps: the corresponding flowchart is reported in Fig. 53.1. In the following subsections, the most important methods currently

Point clouds registration

Point clouds decimation Point clouds decimated

3D Point clouds

Tessellation algorithm

Point clouds filtered

1569

Decimation algorithm

Point clouds filtering

Registration algorithm

Filtering algorithm

3D Point cloud acquisition

Threshold value

3D Metrology for Ancient Pottery Classification and Reconstruction

Threshold value

53

Tessellation

Point cloud registered a)

b) Geometric model editing and exportation

c)

Tessellated surface

Fig. 53.1 Flowchart of the geometric model setup

available are presented for each step, highlighting their merits and defects (for a more informed choice) and reporting, where possible, quantitative values and comparisons.

53.1.1 Point Cloud Acquisition The first step of this process therefore concerns digital data acquisition, which can be performed with different tools. Since the 1990s, 3D scanning technologies have been successfully used in the field of CH ([1–5]) because they have the capability to collect the information needed, allowing scientific studies or fruition of the virtual artifact by the general public, avoiding the risk of possible damage. To date, the most important technologies [6] that meet the technical and economic constraints of 3D metrology for ancient pottery investigation are as follows: – Triangulation laser scanning: the device emits laser pattern over the object, and an optical sensor, calibrated with the laser emitter, identifies the position of this pattern and calculates the depth information by simple triangulation. – Structured light: the scanner consists of a beamer and two cameras; during the digitization process, the projector illuminates the object with patterns of parallel white and black stripes of varying widths. The camera then registers the pattern, taking a picture of each projection. This creates a temporal sequence of different brightness levels used by the system to extract the geometry. – Stereo vision: the system is composed of two calibrated optical sensors. The images are acquired simultaneously, and the distance between corresponding points in both images is used to calculate the depth measurement.

1570

L. Di Angelo et al.

– Photogrammetry: the 3D model is obtained by a collection of images of the object with different points of view; the texture information is also commonly applied together in the process. Table 53.1 reports the main characteristics of the 3D scanning technologies for ancient pottery acquisition. All these optical-based scanning technologies have the limitation of acquiring only the visible surfaces; small undercuts and concave features, for example, cannot be acquired, not being accessible by the sensor. Scanning a whole sherd implies that the object and the sensor must take on different spatial configurations over time. This slows down the process of generating the geometric model and worsens its quality (see following sections). For this purpose, the scanners that implement the technologies presented in Table 53.1 have different configurations and/or devices; the most important ones are as follows: – The 3D scanner is mounted on the end of an anthropomorphic or robotic arm. – The 3D scanner is located on a device in which an accelerometer, a gyroscope, and a compass are mounted; this allows the software to understand the position and movement of the scanner in real time. – The 3D object is located on an automatic rotating table. Furthermore, some devices couple the precision of a laser sensor to acquire geometry with optical systems to map the texture. The choice of the 3D digitalization system is crucial as it directly affects the process time and the quality of the point cloud, which determines the final digital model. Although there are many devices based on different technologies on the market [6], to the present day, no 3D scanner meets all the technical and economic constraints of CH contexts. At this purpose in [7], the authors proposed an AHP (Analytic Hierarchy Process)-based method for choosing the best 3D scanner for cultural heritage applications. The technical parameters considered for the analysis are as follows:

Table 53.1 Comparison of 3D scanning technologies Technology Triangulation laser scanning

Strength High resolution and accuracy

Structured light

Built-in texture Portability

Stereo vision

Built-in texture Portability Low cost Built-in texture Portability Low cost

Photogrammetry

Weakness Reflectance and transparency of the surface may cause wrong measurements High cost Reflectance and transparency of the surface may cause wrong measurements Sensibility to ambient illumination Low resolution and accuracy

Low resolution and accuracy

53

– – – – – – – –

3D Metrology for Ancient Pottery Classification and Reconstruction

1571

Resolution (δm): the minimum distance between captured points Accuracy: refers to both trueness and precision Acquisition rate: the number of points acquired per second Total hourly cost (THC): [investment cost + operating cost]/(hours per year) Texture acquisition Sensitivity to natural scene lighting Sensitivity to surface conditions User interaction in postprocessing

The method is applied to the choice of the 3D scanner for the acquisition of pottery sherds for two different purposes: – Test case #1: classification on the basis of recognized morphological, geometrical, and dimensional features; δm ¼ 1 mm. – Test case #2: exhibition on a website and utilization in 3D printed copies; δm ¼ 1.5 mm. The analysis is performed by considering the most up-to-date devices for each technology used in this application. Table 53.2 summarizes some of the important technical parameters of the devices here considered. In Fig. 53.2, the final results are summarized in two radar graphs; according to the goal, the best choice for test case #1 is the DTLA (Fig. 53.2a), whereas for test case #2, it is the DPH (Fig. 53.2b).

53.1.2 Point Cloud Filtering Data filtering is subsequently carried out so as to reduce the noise that typically affects any measuring process. Though 3D reconstruction applications have been growing due to the improvements in the 3D scanner technology, there is still another problem: the raw data produced by 3D scanners are noisy and cannot be used directly into the 3D reconstruction process without a previous processing. Typically, it consists of two different phases: – Outliers’ detection and erasing – Smoothing The word “outlier” has been defined in several ways depending on the applications. In simple words, an outlier is an observation that is so far from most others or that in some way differs from the general pattern of the majority, which should be treated differently. The presence of outliers is an unavoidable phenomenon in point cloud data. Outliers occur mainly due to the physical limitations of data collection sensors, discontinuities at boundaries between 3D features, occlusions, multiple reflectance, and noise that produce off-surface points [8]. Since points are usually unorganized, noisy, sparse, and inconsistent in point density or they have

1572

L. Di Angelo et al.

Table 53.2 The devices analyzed with the total hourly costs and the principal technical specifications [7] Technology PH

Model Z Scan Micro

Identifier DPH

Costs THC 5 57.20 €/h

SL

Artec Eva 3D Scanner

DSL,1

THC 5 57.95 €/h

SL

EinScan-S Desktop 3D Scanner

DSL,2

THC 5 26.83 €/h

TL

NextEngine 3D Scanner Ultra HD

DTL

THC 5 57.27 €/h

TL A

Faro Arm 2.7 m Edge

DTLA

THC 5 59.67 €/h

DTLA

Technical specification Resolution: 0.25 mm Accuracy: 0.05 mm Acquisition rate: 18 kpoints/s Resolution: 0.5 mm Accuracy: 0.1 mm Acquisition rate: 2 Mpoints/s Resolution: 1.3 mm Accuracy: 0.1 mm Acquisition rate: 50 kpoints/s Resolution: 0.127 mm Accuracy: 0.1 mm (macro); 0.4 mm (wide) Acquisition rate: 50 kpoints/s Resolution: 0.040 mm Accuracy: 0.025 mm Acquisition rate: 560 kpoints/s

DTLA

37,15%

DPH DPH

DTL

DTL 25,70%

19,82%

17,32%

a) DSL1

b)

DSL2

Fig. 53.2 Selection results for test case #1 (a) and test case #2 (b)

DSL1

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1573

geometrical discontinuities or arbitrary surface shape with sharp features, outlier detection in point cloud data is still an open issue. Existing methods for outlier detection can be broadly arranged into four groups: – Statistical methods: where statistical techniques fit a statistical model (usually for normal data) to the given data and then apply a statistical inference test to determine whether an incoming instance fits the model or not ([9, 10]). – Distance and/or density-based methods: local outliers are identified by examining the distances to the nearest neighbors [11]. – Clustering-based methods: they group similar data instances into clusters and consider clusters of small size as outliers [12]. – Model-based approach: they are based on the learning of a distinctive model from a set of training data instances and detect outliers as deviations from the model [13]. Table 53.3 reports the most important limitations of the methods for outliers’ detection. Hodges and Austin, in [14], concluded that there is no single universally applicable or generic outlier detection approach, but people are trying to get more effective methods based on various applications. The development of robust point cloud smoothing algorithms has received much attention in last years. A complete review is published in [15]. According to this work, the proposed algorithms can be categorized into the following seven groups: – Statistical-based filtering techniques: they utilize the adaptation of the statistical concepts, which are suitable for the nature of the point cloud. – Neighborhood-based filtering techniques: they determine the filtered position of a point using similarity measures (positions of points or normals) between a point and its neighborhood. – Projection-based filtering approaches: they adjust the position of each point in a point cloud via different projection strategies. – Signal processing-based method: typical methods for signal processing, such as the Laplacian operators, discrete approximation of Laplacian, and Fourier transformation, are applied to smooth the mesh. Table 53.3 Comparison of outliers’ detection methods Outliers’ detection methods Statistical methods Distance and/or densitybased methods Clustering-based methods

Model-based approach

Weakness Information about the underlying data distribution may not always be available Results depend on parameter k used to define k-nearest or k- distance neighborhood to compute the outlier factors Performance is highly sensitive to the clustering techniques that are involved in capturing the cluster structure of the regular (inlier) data Inefficient at detecting small planes in large point clouds

1574

L. Di Angelo et al.

– PDEs-based filtering technique: the position of the smoothed points is obtained by computing the partial differential properties. – Hybrid filtering techniques: they use two or more smoothing techniques together to deal with the raw point clouds. – Voxel grid-based method: the smoothing is performed by approximating all the points that lie in a voxel with a properly chosen point. The results reported by the authors can be summarized as follows: – Voxel grid-based method can be recommended as a better choice regarding computational efficiency, but its filtering effect is unsatisfactory. – Statistical-based filtering techniques typically demonstrate a sufficient performance in terms of noise removal, together with feature preserving, but they are computationally expensive. – Projection-based filtering approaches can be considered as a relatively good trade-off, which provides a balance between running efficiency and the filtering effectiveness. In a specific application for archaeology, in [14], the data are smoothed by using a Gaussian filter (statistical-based filtering technique) implemented in Geomagic ® and set to produce a mean displacement of the points not exceeding 0.02 mm.

53.1.3 Point Cloud Registration and Decimation Since an archaeological find cannot be acquired with a single scan, it is necessary to capture various views of the object. To obtain a single and coherent point cloud, the different acquisitions have to be combined together under the same coordinate system by a process called registration. Typically, in the cultural heritage applications, the registration is performed two clouds at a time, according to the following two main steps: – A preliminary rough registration carried out by the user which selects three or more correspondent points on the two different clouds. – A fine registration by means of the iterative closest point (ICP) algorithm [16], which, iteratively, searches the configuration minimizing the sum of the distances of the points of the two clouds. In [17], the authors assumed the registration has been performed successfully when the mean distance between the two point clouds is lower than 0.08 mm. The point cloud registration process is a heavily studied problem in computer vision. Nevertheless, several open questions continue to emerge; one of the most important is the registration of more than two point clouds. In this case, by using the strategy presented above, due to error accumulation and propagation, the registered point clouds have gaps. The methods proposed in the related literature [18], and not

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1575

yet implemented in any commercial software, can be used starting from an initial alignment. Once the two point clouds are registered, a uniform point sampling is performed in order to make the point density uniform. Depending on the size of the features to be analyzed, point sampling may vary from 0.15 mm [17] to 1 mm [19].

53.1.4 Point Cloud Tessellation With the aim to obtain an oriented surface, the points are interpolated by triangular planar facets. Various algorithms to tessellate point clouds have been proposed in the related literature; the most important ones are generally divided into three main categories: – Implicit methods: the tessellation is obtained by extracting the nominally null level of an implicit function f(p) ¼ 0 (where p denotes either all point cloud or only a part of it), formulated so as to be negative inside the point cloud and positive outside. The published methods differ for the formulation of implicit function ([20–24]). – Delaunay-based methods: they generate surfaces starting from a volume tetrahedralization by means of a 3D Delaunay triangulation of the sample points. The most important methods presented in the related literature ([25–27]) differ essentially in the way they remove the tetrahedral and build the external triangular mesh. – Mesh growing-based methods: the surface reconstruction starts with a seed triangle, and the meshed area is grown by pushing the fronts ahead using some criteria ([28–32]). Table 53.4 summarizes the advantages and drawbacks of the three categories of tessellation methods. In [31], the authors compare the performance of the following methods in the tessellation of some benchmark point clouds and artificially noised test cases:

Table 53.4 Comparison of tessellation methods Tessellation methods Implicit

Delaunay-based Mesh growing-based

Strength Closed textured surface even in case of scattered and noisy data Theoretical guarantee of a good reconstruction High-speed tessellation

Weakness Long processing times Since the surface does not pass through all the points, there may be a loss of detail Only for closed surface Low-speed tessellation There is no theoretical guarantee of a good reconstruction

1576

L. Di Angelo et al. G2S_OLD

G2S

BPA

COCONE FAMILY

RATE 5 HOLES

OUTLIERS ADDED POINT CLOUD

NON-MANIFOLDNESS

4 3

CLOSED SURFACES

2 NON-MANIFOLDNESS

HOLES

1 0

NON-MANIFOLDNESS

HOLES NOISE ADDED POINT CLOUD

NON-MANIFOLDNESS

HOLES

OPEN SURFACES

Fig. 53.3 Comparison results of some tessellation methods [31]

– Ball pivoting method (BPA) [28]: mesh growing-based approach for which a new triangle is constructed for each front edge by connecting its extreme points with the point touched by a ball (of user-defined radius) pivoting around that edge. – Cocone family ([25, 26]): a Delaunay-based approach where candidate triangles are those inside the complement of a double cone of assigned opening angle and whose apex is the point under analysis (p), and the axis is the normal at p. – Gabriel 2 – Simplex (G2S_old) [31]: mesh growing-based approach centered on the following criterion: a triangle is a G2S if its smallest circumscribing ball is empty. – Robust Gabriel 2 – Simplex (G2S) [32]: mesh growing-based approach that is an improvement of the G2S_old by introducing an original priority queue for the driving of the front growth and a post processing to efficiently erase the non-manifold vertices. Figure 53.3 summarizes the results: the methods are evaluated by assigning a score 1–5.

53.1.5 Geometric Model Editing and Exportation Depending on the quality of the point cloud and the method used for the tessellation, the geometric model may have defects; the most common are (Fig. 53.4): – Vertices (highlighted in red in Fig. 53.4a) and/or edges not manifold (highlighted in red in Fig. 53.4b) – Holes (Fig. 53.4c) – Not congruent normals (Fig. 53.4d) – Self-intersecting triangles (Fig. 53.4e) In order to analyze the geometric model, these defects must be eliminated. For this purpose, the software commonly used have specific tools.

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1577

v

a)

c)

b)

d)

e)

Fig. 53.4 Typical defectiveness generated by tessellation methods

Once a manifold discrete model of the original find is obtained, it is exported in an STL format. The file represents a solid whose surface has been discretized into triangles. It results from the X, Y, and Z coordinates repeated for each of the three vertices of each triangle, with a vector to describe the orientation of the normal to the surface. This type of format file has two formats: the ASCII and binary format that is the more efficient in terms of memory usage (it is one-sixth the size of the ASCII format).

53.2

Geometric Model Fragment Processing

The geometric model of the pottery needs to be processed in order to get the most information. The latter consists of geometrical, morphological, and dimensional characteristic features of ceramics that contribute to analyze and eventually classify pottery’s sherds. The geometric model processing belongs to three different areas: – Axis identification – Feature segmentation and recognition – Dimensional feature evaluation The first phase is focused on axis identification. Pottery is substantially characterized by an axially symmetric geometry, so axis identification is a preliminary step to take before dimensional analysis of pottery or of its fragments. The second phase identifies fundamental features of fragments: the geometrical and morphological ones. Morphological features typify the not axially symmetric part of fragments (such as handles, ribs, and decorations). Geometrical features pertain to the axially symmetric part of pottery (internal/external wall, rim, base). Lastly, in the third area, the processing step regards the dimensions that characterize the geometric and morphological features pertaining to the sherd’s pottery.

1578

L. Di Angelo et al.

The processing of a fragment’s model allows exploiting the abovementioned information, even from non-indicative fragments. This type of fragments usually pertains to the wall section of the pottery, and it is difficult to analyze following the traditional method used in archaeology.

53.2.1 Axis Identification The axis evaluation is a fundamental issue in the archeological field since it is used for the analysis of the fragment, which is at the base of the computer-based systems developed for ancient pottery classification, and for retrieving the original pottery shape. Necessarily, then, the axis estimation has to be very accurate because it affects the final classification and the archiving of the data concerning the finds. Since the axis of symmetry is a nonphysical geometric entity, it cannot be directly extracted by the measured surface of an axially symmetric object. Therefore, the axis has to be derived taking advantage of some properties of axially symmetric surface, from which 3D high-density point cloud is extracted. Detecting the axis from discrete geometric models is a difficult task, and in the case of archaeological finds, it is even more complicated due to several reasons. Firstly, the geometry of an archaeological vessel is only approximately axisymmetric since it may include some features, such as handles, decorations, inscriptions, lugs, feet, and, for the specific case of a sherd, fracture surfaces. The evaluation of the axis, therefore, requires as a preliminary step the removal of these features, and this is generally done in a single step, recognizing the axisymmetric portion of the vessel or fragment surface. Secondly, the axially symmetric surface of an archaeological pot is generally affected by several imperfections due, for example, to extensive wear, encrustations, chipping and weathering, and other damage, so its sections are not circular and are not concentric with each other. These factors inevitably influence the quality of the axis estimation and consequently the quality of the measurements and evaluations that derive from it. The published automatic methods can be grouped as follows: – Normal-based methods ([33]), which are based on the property of axially symmetric surfaces for which the normal vectors at any point of the surface always intersect the axis of symmetry. – The osculating sphere-based method [34] exploits that the center of the maximal sphere tangent to the surface (osculating sphere), evaluated at each node of the tessellated surface, is located on the axis of the axially symmetrical surface. – The symmetry line-based method [35], which is based on the property for which the symmetry line of any curve resulting from the planar section of a complete axially symmetrical surface always intersects the surface axis. – The circle and line fitting methods ([36–39]) exploit the property whereby each section of an axisymmetric surface perpendicular to the axis is a circle, whose center is on the axis. An important difference among the implemented methods is

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1579

that those proposed by Mara [36] and Son [39] estimate the axis without the evaluation of the differential geometrical properties. – The thickness versor intersection-based method [40] is based on the property the minimum wall thickness line always intersects the axis of revolution. The results of the experimentations and comparisons performed in [40, 41] of the abovementioned methods’ performances in the axis detection from several axially symmetrical geometries (analytical and generic, partial and complete, synthetic and real) for different types of discretization (sample rate and regularity of the mesh) are summarized in Table 53.5. Table 53.5 Comparison of axis detection methods Tessellation methods Normal-based

Strength It provides a quick estimation without requiring any preliminary evaluation

Osculating sphere

Typically, it provides a better solution compared with the normal-based method

Symmetry line-based

It provides a robust and accurate axis estimation also in the presence of singularities and noise Recommended for metrological applications

Circle and line fitting

Thickness versor intersectionbased

The method seems to be robust in the axis identification also on an object affected by extensive wear, encrustations, and all the typical damage that can be found in a common archaeological sherd

Weakness Sensitive to the measurement noise, outliers, surface roughness, and all those factors that make the surface not perfectly axially symmetric Sensitive to the measurement noise, outliers, surface roughness, and all those factors that make the surface not perfectly axially symmetric A preliminary axis estimation is required It fails when the fragment’s shape is close to spherical Only for complete axially symmetric surfaces It not able to provide accurate results for objects with incomplete circular span, especially when affected by micro and macro irregularities (such as roughness and local deformations) that make the surface not perfectly axially symmetric (such as in the case of fragments of archaeological vessels) It fails when the fragment’s shape is close to spherical Only for thin-walled axially symmetric surfaces

1580

L. Di Angelo et al.

53.2.2 Profile Evaluation The geometry of an axially symmetric sherd is defined by the pair axis (α)/profile [42]. An axially symmetric surface, in fact, can be always generated by revolving a planar curve, named profile, about the axis of rotation (or axis of symmetry). For a sherd, the profile is a closed curve (Fig. 53.6). The profile extraction is generally the first step in documentation, classification, and reconstruction of archaeological ceramics. The profile, in fact, holds most of the information needed to perform these tasks. Traditional methods for extracting the profile are based on hand-drawn sketches, which always result from a subjective interpretation of the sherd geometry: the draftsman catches the representative shape of the profile by removing everything considered poorly characteristic of the overall fragment geometry. These sketches are often inaccurate and cannot handle the volume of information held into the fragment within reasonable time and costs. For the large-scale documentation of ceramic fragments, the outline of the sherd profile is often obtained by a profile gauge. Several computer-aided and automatic methods were proposed for profile extraction ([17, 43–45]). The underlying principle of these methods is that starting from an axial symmetric surface, the profile can be obtained by sectioning the surface with any half-plane passing through the rotation axis. The profile thus established is, however, uniquely determined only in the case of an ideal surface. For an archaeological fragment, the resulting profile depends on the plane due to local surface imperfections (deformations, roughness, etc.) generally affecting the sherd surface. Therefore, for automatic methods, the first question to be addressed concerns which of the profiles should be used in order to represent a given fragment. The most important methods published in the related literature differ from how the mean profile is determined. Typically, the state-of-the-art approaches consist of two main steps: a first attempt profile identification and the final representative profile. The exception is the approach proposed by Kampel et al. in [44] that identifies the representative profile of a sherd as that in correspondence of fragment maximum height. In Table 53.6, the most important characteristics of analyzed methods are summarized.

53.2.3 Feature Detection The aim of an effective computer-based method is the automatic recognition of fragment features in accordance with the way archeologists perform the conceptual categorization of ceramic. The translation of the traditional way of studying archaeological pottery in an automatic method is not a trivial process. This is because the characteristic elements of sherds are not associable to analytical surfaces; so specific operative rules, which are valid for the abovementioned feature recognition, are required for the widest set of types of pottery fragments. Moreover, this process is

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1581

Table 53.6 Comparison of profile detection methods First attempt profile Method identification Hlavackova- A section of the Schindler surface with a halfet al. [43] plane passing through α

The final representative profile The profile is divided in several parts, and each of them is approximated by a cubic B-spline curve

Karasik and Projection on the Smilansky plane z-ρ (z  α and [45] ρ is the distance from α) of all the points of the surface (except those recognized belonging to fractured surface)

Di Angelo et al. [17]

Kampel et al. [44]

The center of mass of the points within rectangles with width inversely proportional to the mean curvature at the point and oriented perpendicularly to the line locally approximating the profile Points smoothed by The radii of the circles centered on a parabola α approximating the approximating the slices perpendicular neighborhood of to α each point profile with a defined width

Profile in correspondence of fragment maximum height

Strength Universal approximation property

Weakness Relatively simple mathematical definition The number and complexity of the curves approximating the profile are not known a priori Procedure valid The final result, of each type of depending on the profile width of the rectangles (calculated using the differential geometrical properties), is affected by the noise of the profile

Ease of implementation and speed of calculation A valid procedure for each type of profile Ease of implementation and speed of calculation

The optimal dimensions of neighborhood are not known in advance

Evaluation affected by local and global surface imperfections

made more difficult since the 3D discrete models come from finds that are generally damaged and worn, so their geometric properties have been lost. For all these reasons, in the last years, the automatic methods published in the related literature have proposed new shape descriptors that, on the one hand, were easily calculable, but for which, on the other hand, there was no obvious correspondence with those used by archaeologists to study finds ([36, 44–47]). Only recently the research group headed by Professor P. Di Stefano of the University of L’Aquila published (in [17, 48]) an automatic method matching the previously mentioned requirements. Based on the analysis of the differential geometrical properties and topological invariants, the method performs the feature recognition through the steps briefly described in the flowchart of Fig. 53.5.

1582

L. Di Angelo et al.

Discrete valid model of sherd

Axis detection algorithm

Axially-symmetric feature recognition Non-Axiallysymmetric surface

Axially-symmetric surface

Morphological features recognition

Geometrical features recognition

A

Morphological features

Internal wall

Fractured surfaces

External wall

Chips

Rim

A

View A Base

View A

Fig. 53.5 Flowchart of the feature recognition method proposed in [17, 48]

The method applied by the authors to hundreds of real archaeological fragments correctly recognizes all the features defined here, in the most cases; in some cases, the method fails in the identification of: – Chips (little isles encircled by axially symmetric regions) located between two different pottery features – Fractured surface when some encrustations are present – Boundary between adjacent features when it is blurred by damages and wear

53.2.4 Secondary Feature Detection Ancient pottery often shows some detail features, not necessarily axially symmetric, that are significant for historical and archeological investigation and for classification purposes. These features result in traces on potsherd surface, which can be left by finger action or by tools both intentionally, such as inscriptions and decorative motifs, or unintentionally, such as working marks. From a geometric point of view, these detail features are obtained by a sweeping action that leaves negative or positive traces on ceramic surface. These features are characterized by a cross section approximated to one or more circular arcs and may belong to the axially symmetric portion of a fragment. Some methods were recently proposed for their automatic detection and measurement, starting from high-density tessellated models, acquired by 3D scanning. These algorithms are based on the processing of geometrical differential properties. The automatic recognition of these detail features from discrete model is an extremely difficult task for several reasons. The high-density tessellated surface is not differentiable, so curvature cannot be evaluated, but it must be recognized from it. These detail features, moreover, are of limited extension with respect to the mesh sampling, so the evaluation of geometrical differential

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1583

properties is more difficult and affected by greater uncertainties [49]. Furthermore, in the case of archaeological artifacts, the evaluation of differential properties is blurred by measurement errors and surface irregularities, such as working marks and surface roughness that induce small fluctuations of the surface curvature. In [50], Solbrig et al. proposed a method that detects the ridges and valleys of triangle meshes of decorated potsherds by finding the local minimum and maximum of principal curvatures. These ridges are detected in correspondence of convex ornamentations, such as cordons, pointed lugs, and embossed decorations. Valley lines are identified from concave ornamentations, such as grooves, channels, engraved lines, and impressions. Ridge and valley lines detection, however, does not allow obtaining information on shape and dimensions of these features that is an important aspect to consider for classification purposes or for identifying the tools used to create them. In [19], a computer-based methodology, aimed at the automatic detection and measurement of decorations with constant radius in archaeological pottery, is proposed. The methodology, conceived under the hypothesis that the values of radii are a priori unknown, consists of two main steps: – Constant radius feature segmentation, which is performed by means of a fuzzification of principal curvatures at the nodes of discrete geometric models – Radii measurement The results reported in [19] show the method is suited to analyze decorations of real 3D scanned objects both negative, such as engraving, graffiti, excised decoration, impression/stamping, burnishing, and roulette, and positive, such as barbotine, applied/plastic, and molded relief.

53.2.5 Dimensional Analysis of Fragments Typically, dimensional features used by the archaeologists to drive the sherd classification are listed in Fig. 53.6. To the best of the authors’ knowledge, Di Angelo et al. propose the only automatic method, published in literature, suited to evaluate the abovementioned dimensional features of fragment in [14]. Based on feature segmentation and recognition method detailed in [48], the proposed approach consists of the steps summarized in the flowchart of Fig. 53.7. In order to verify the method robustness, the authors compare the characteristic diameters of a modern common pottery with the corresponding values obtained after its breaking into fragments. The low percentage differences prove that the method is suitable both to recognize fragments pertaining to the same pottery and to draw correct dimensional information of the pottery from its sherds. Furthermore, the percentage errors are strictly affected by point density values (greater than 1 mm) and by the method used for axis detection.

1584

L. Di Angelo et al.

rim -

internal wall

φR

base

external wall

A tR1

φN

(A, B): endpoints situated at the base and rim (M, N): points of maximum/minimum

φM

M

tw

tB

diameter where the tangent is vertical

tR2

N

the thickness of the rim (tR1); the height of the rim (tR2); the diameter of the pottery at the apical point of the rim (φR); the maximum diameter of convex part (φM); the minimum diameter of a concave surface (φN); the thickness in correspondence of the base (tB); the maximum diameter of the base of the pottery (φB); sherd height (hp); the thickness of the wall (tw).

hp

B

φB

representative sherd profile evaluation

A

View A

approximation of the sherd profile

segmentation of the external part of the profile

smoothing of the sherd profile smoothing algorithm

alignment of axis α of the sherd to the z-axis

profile segmentation

segmentation algorithm

approximation algorithm

3D model of the sherd with features segmentation and recognition

segmentation algorithm

Fig. 53.6 Dimensional features and characteristic points

concave and convex tracts segmentation of the external part of the profile

Dimensional features evaluation φN,2 92.3 φM,3 100.5

φB 95.9

Fig. 53.7 Dimensional fragment analysis [48]

53.3

Geometric Model Processing of Whole-Shape Pottery

With regards to whole-shape pottery, the complete analysis can be performed analyses by the methods and evaluations presented in the previous paragraphs. Since the whole geometric shape is available, the evaluations are more accurate and robust. A specific evaluation of whole-shape pottery is the fitting volume. Ceramics in archeology, in fact, have a fundamental implication in the reconstruction of ancient trade and economy in general. Therefore, quantifying the contents of the vessels

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1585

through volume and weight evaluation is helpful to this end: typological and metrological pottery standardizations could be verified from such analyses. Traditionally, it can be done with a liquid’s displacement from a measuring jug. Anyway, by dipping vessels in water, delicate materials and eventual figurative decorations can be ruined. In literature, the purpose of automatically estimating the volume of ancient vases is done through 2D profiles of ceramics ([51, 52]). These methods are directly applicable in the case of 3D scanned models. Difficulties are in the evaluation of the internal profile of closed-shape ceramics, as they are difficult to measure, also with modern 3D scanners (laser or structured light); this is especially for objects with narrow opening. Although the computer tomography (CT) scan results in accurate data also of interior surfaces, this technology is not used since it is expensive, and the used radiation might interfere further examinations (radiocarbon dating). This problem cannot be solved by using the published methods to generate the offset of the external surface without overlapping in geometry ([53, 54]); in fact, the existing interior part is not taken into account. In order to overcome this problem, in [55], the authors proposed a technique for reconstructing the inner surface of the pottery by using the incomplete scanned inner geometry. The calculation is done iteratively to minimize the difference between expected and achieved object volume. The reported results show the method generates a fully reconstructed, closed manifold mesh without self-intersections or overlaps.

53.4

3D Information Systems for Archaeological Pottery Visualization and Documentation

The storage and retrieval of archaeological data within computer databases are a basic component of modern archaeological research [56]. The ceramic databases try to standardize many repeated tasks encountered during the recording of archaeological and ceramic data in an efficient framework. They are a support for classification and interpretation, especially in post-excavation analyses. Creating these databases is not a trivial activity since they have to manage very heterogeneous set of sources, data structures, content, and formats. To date, relational database management system (RDBMS) is often used in archaeology to register and to document archaeological excavations, including information that describes the historical, stratigraphic, urban topographic context, with a focus on pottery. The available databases relative to archeological ceramic do not have a common structure and format; consequently, the exchange of data among researchers is difficult. By analyzing the published databases/structures, relative to the shape and morphology of sherds, it results they report digitization of out-of-print fascicules ([57, 58]) or information performed by skilled operators [59]. In Table 53.7, the most important available databases/structures, which start from a 3D scanned model of sherd, are listed and compared in the voices outlined. Furthermore, annotations about quantity (number of sherds), origin, class, shape, typology, decoration, production, chronology, and bibliography can be added to the

1586

L. Di Angelo et al.

Table 53.7 Comparison of 3D information systems for archaeological pottery Structure/databases The Pottery Informatics Query Database (PIQD) [60]

Computer-Assisted Drawing of Archaeological Pottery (DACORD) [61]

Voices outlined The morphology of the ceramic vessel (coming from both 2D scans of illustrated ceramics profile and 3D acquisitions) is converted in measurements of radii, tangent, and curvature of its representative profile as a function of arc length. General layout (layout is represented in the plane y-z, where z-axis coincides with the axis of symmetry) Overall form (the selected profile section contains information for further fragment classification, e.g., rim, foot, handles, distance measurements, volume estimation, maximum sector angle) Fragment rendering Vessel regularity (evaluated as distance between model points and axis or model points and the 3D object obtained by rotating the profile around the axis) 3D reconstruction (the 3D object obtained by rotating the profile around the axis) Illustration exportation (archeological illustrations, 3D reconstruction)

structure. The construction and updating of the abovementioned structures are not completely automatic; the papers [60, 61] refer to features (rim, foot, characteristic dimensions) without presenting an automatic method for their identification, especially for fragments.

53.5

3D Puzzling of Archaeological Fragments

Archaeological ceramics are often found in the form of fragments, so the problem of their reassembly, in order to obtain a reconstruction of the ancient artifact, is of great interest to achieve a typology and a final classification of the find. Reassembling ancient pottery from fragments is, however, a complex issue due to several complications, such as [62]: – Physical degradation of fragments, due to chipping and erosion (especially if exposed to weathering) – Unknown number of pots to be reconstructed when fragments come from a collection or context not well documented – Missing fragments, because they are destroyed or not yet discovered The approach, generally used for pottery reconstruction, involves two main tasks: – Clustering of matching candidate fragments – Fragment assembly

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1587

The archaeologists generally identify the fragments belonging to the same ceramic based on morphological aspects. Two are the main approaches used to take into account the sherd morphology: the descriptive and the metric. The descriptive approach aims to characterize the overall shape of the artifact by comparing it with familiar templates. The metric language tries to quantify the morphology by a set of local and discrete metric measures, which, in the case of the traditional manual methods, are taken by sliding caliper, profile comb tool, circle template, etc. [63]. The geometric shape is not the only factor to consider in assessing the similarity between fragments; other aspects, such as decorative motifs, color, textures, material, chemical properties of color pigments, and many more, contribute to characterizing the sherd’s morphology [64]. The traditional methods, which require the manual effort by expert ceramics analysts, are laborious, time-consuming, and expensive. Other important shortcomings of these methods are the poor accuracy and repeatability of metric evaluations, so the results of the reconstruction depend on the ability of the expert performing it. Objectifying and speeding up this process is therefore important, and recently, several computer-aided methodologies, automatic or not, were proposed to support the archaeologists of abovementioned two complex tasks. When a bunch of sherds from different pots are mixed together, it is necessary to separate them before pot assembly. Several computerized methods have been developed to help archaeologists. Han and Hahn in [38] proposed the distribution of principal curvatures for grouping sherds, preliminarily clustered by an operator according to their location in a pot, with similar color and texture. A more general method is proposed by Biasotti et al. in [64], in which the similarity is characterized by the following criteria translated into computational tools: overall fragment size, thickness, material texture, shape continuity, color, and decorations. The application of these tools to real cases, however, shows that they have the potential to support archaeologists with quantitative estimates of their findings, but they are still far from giving a satisfactory solution to the problem of assessing similarity. This is due to several problems, which span from the need of keeping results consistent across different resolutions to the management of damaged fragments. The reassembly of an arbitrary object from fragments, known in literature as mosaicing, can be generally viewed as solving a 3D puzzle. In [65, 66], the authors proposed graphics interfaces to help the user to assemble manually fragments on a computer. The published methods to automatically reassemble the pottery fragments can be grouped as follows: – Global methods: they assume that the overall shape of the broken object is known a priori. – Local methods: they rely on accurately aligning only two fragments at a time and the reassembly of the object is obtained incrementally. In the global methods [67], the orientation and position of the fragments along the axis α is defined by their matching with the geometry of the template of the axially symmetric pottery; their angular positions are determined by matching the contour curves and fractured surfaces of overlapping fragments.

1588

L. Di Angelo et al.

Table 53.8 Comparison of methods for sherd reassembly Reassembly methods Global methods Local methods

Strength Simplification of the process of aligning each sherd It is more general

Weakness The shape of pottery has to be known The alignment process can be laborious (hundreds of possible configurations) Accumulation of errors Strongly affected by the physical degradation of fractured surfaces

The local approaches consider sherd pairs as the basic building block of the assembly algorithm. Known the axis α of the sherds, the searching problem for fragment matching is solved by fixing a sherd and finding the roto-translation (rotation around α and translation along α) transformation of the other one which minimizes an objective function. Typically, matching sherds include the contour curve and fractured surfaces matching ([68, 69]). Willis and Cooper in [70] proposed to minimize a distance function of point coordinates and normal of contour curves. Table 53.8 summarizes the strength and weakness of two general strategies. Both approaches can provide a global relaxation ([71]) in order to minimize simultaneously all local alignment errors.

53.6

Classification

Ceramics’ classification is a significant and complex operation that addresses to group similar objects from an archaeological excavation or a surface collection based on some attributes [72]; in such a way, it is possible highlighting the relationships between them and possibly reciprocal dependencies. In archeology, the ideal classification is the one that relates the set of pots manufactured in a given geographical area, in a precise period of time and with similar functions. In setting a classification/ typology, it is essential to choose the units that make up its base and on which the classification itself is built. The archaeologists have mostly worked on the creation of a general vocabulary to be used in classification, and relatively little attention has been paid to obtain a relevant classification method focused on the problem of replicability of observational units [73]. As a result, a wide variety of terms and words describing features has been created among archaeological studies and works. Traditionally, the classification of ceramic material is based on shape, dimensions, decoration, technological elements, color, and material. Nowadays, all of these features are recognized and analyzed by a skilled operator. Some of them, such as decoration, technological elements, and color, are investigated by a visual analysis. The shape and dimension characterization are largely based on the graphical representation of findings [74]. All this information is reported on printed catalogues; the operator, by visual comparison with known classifications, assigns to a specific ceramic typology each analyzed sherd. Since ceramic materials are typically found

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1589

in fragments, the previous operations are more complex to be performed. In a previous paper [48], the authors demonstrated that the traditional method for shape and dimensional characterization is not reproducible and repeatable. Furthermore, the manual method has the significant limitation of making it possible the orientation of only fragments (referred to as “indicative”) having features such as pieces of rim, base, presence of grooves; this is because without these features, it is not possible to adequately estimate the axis of the sherd. As consequence, the most part of finds (walls without pieces of rim, base, or groove) that has important information about the shape of the pottery original profile is not analyzed. In order to speed up, objectify, and extend observational units’ evaluation and comparison, some automatic or semi-automatic approaches are proposed. The goal of such research topic is to reproduce the archaeologists’ work method, which consists of analyzing the specific characteristics of pottery and then comparing them in a well-known typology. According to Karasik and Smilansky [75], any computer-based classification method should satisfy the following mandatory requirements: – – – – – – – –

The procedure should be efficient and cost-effective. It should enable hierarchical storage and retrieval of large data. The archaeologist should be able to use the algorithm interactively. The classification should use the information stored in the entire profile of the fragment and not only in discrete metrical measures. The procedure should be able to compare fragments of different sizes. The classification should be hierarchical and flexible. The classification scheme should be amenable to generalizations. The resulting classification should make archaeological sense and pass the scrutiny of expert ceramics analysts.

The computer-based methods proposed in the related literature perform the classification by considering the shape and/or the dimensions; they can be grouped into three principal categories: – Segmentation-based ([76–78]): these methods first perform profile identification, then its segmentation into 2D geometric features. Evidence of pottery classification comes from the relative distances between the recognized points. – Shape descriptor-based ([75, 79–83]): the pottery characterization is carried out with one or more shape descriptors of the 2D profile and with the presence of appendages (typically, handles and supports). – Direct matching ([84–88]): the classification is based on the direct comparison of the obtained profile with reference ones. The automatic classification of archeological pottery is still an open issue, since none of the analyzed methods verify all mandatory requirements proposed by Karasik and Smilansky [75]. For this purpose, two European funded projects are proposed:

1590

L. Di Angelo et al.

– The Gravitate project (https://gravitate-project.eu/): ended at November 2018. It was focused on creating a set of software tools for experts in the CH field to reconstruct fragmented cultural objects, to identify and re-unify parts of a cultural object once belonging to the same collections, and to recognize associations between cultural artifacts that will allow new knowledge and understanding of past societies to be inferred. – The Archaide project (http://www.archaide.eu/): it aims to create a new system for the automatic recognition of archaeological pottery from excavations around the world by developing new apps based on latest automatic image recognition technology.

53.7

Additive Manufacturing Technologies for Physical Reconstruction of Ceramics

The additive manufacturing of 3D digital models of artifacts allows to introduce new revolutionary approaches for the cultural heritage world: virtual museum creation, artifact cataloguing, conservation, planning and simulation of restoration, monitoring of artifacts subjected to environmental degradation, virtual reconstruction of damaged or missing parts, and replica or souvenir reproduction. Regarding ceramics, additive manufacturing techniques have opened up new opportunities ([2, 89, 90]) such as: – Setting up tactile paths for the blind and visually impaired, supporting their experiences in museums and CH in general. 3D replicas allow exploring sculptures or artworks with fingers, without getting in direct contact with the original. – Production of tailored packaging for supporting, displaying, and shipping cultural objects. – Temporary or permanent replacement/loans of originals for exhibitions; replicas would be useful to the repatriation of original artifacts (as opposed to what happens today, replicas would be kept in the possession of the organization, while the original artifact would be returned to its possessor). – Engaging audiences with archaeological processes, such as reconstructing a shape from a given group of shards or pieces. – Commercial production of physical copies.

References 1. Pieraccini M, Guidi G, Atzeni C (2001) 3D digitizing of cultural heritage. J Cult Herit 2(1): 63–70 2. Segreto T, Bottillo A, Teti R, Galantucci LM, Lavecchia F, Galantucci MB (2017) Non-contact reverse engineering modeling for additive manufacturing of down scaled cultural artefacts. Procedia CIRP 62:481–486

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1591

3. Cheng H-M, Yen Y-N, Yang W-B (2009) Digital archiving in cultural heritage preservation. In: Proceedings of the 14th conference on computer-aided architectural design research in Asia, CAADRIA2009, Yunlin, Taiwan, pp 93–101 4. Hassani F (2015) Documentation of cultural heritage techniques, potentials and constraints. Int Arch Photogramm Remote Sens Spat Inf Sci 40(5W7):207–214 5. Levoy M et al (2000) The digital Michelangelo Project. In: Proceedings of the 27th annual conference on Computer graphics and interactive techniques, pp 131–144 6. Weigert A, Dhanda A, Cano J, Bayod C, Fai S, Santana Quintero M (2019) A review of recording technologies for digital fabrication in heritage conservation. Int Arch Photogramm Remote Sens Spatial Inf Sci XLII-2(W9):773–778. https://doi.org/10.5194/isprs-archivesXLII-2-W9-773-2019 7. Di Angelo L, Di Stefano P, Fratocchi L, Marzola A (2018) An AHP-based method for choosing the best 3D scanner for cultural heritage applications. J Cult Herit 34:109–115 8. Nurunnabi A, West G, Belton D (2015) Outlier detection and robust normal-curvature estimation in mobile laser scanning 3D point cloud data. Pattern Recogn 48(4):1404–1419 9. Barnett V, Lewis T (1995) Outlier in statistical data. Wiley, New York 10. Rousseeuw PJ, Leroy A (2003) Robust regression and outlier detection. Wiley, NewYork 11. Knorr EM, Ng RT (1998) Algorithms for mining distance-based outliers in large datasets. In: Proceedings of the 24th international conference on very large databases (VLDB), NewYork, USA, pp 392–403 12. Jiang SY, An QB (2008) Clustering-based outlier detection method. In: 2008 IEEE Fifth international conference on Fuzzy Systems and Knowledge Discovery 2:429–433 13. Liu B, Xiao Y, Cao L, Hao Z, Deng F (2013) SVDD-based outlier detection on uncertain data. Knowl Inf Syst 34(3):597–618 14. Hodge V, Austin J (2004) A survey of outlier detection methodologies. Artif Intell Rev 22(2): 85–126 15. Han XF, Jin JS, Wang MJ, Jiang W, Gao L, Xiao L (2017) A review of algorithms for filtering the 3D point cloud. Signal Process Image Commun 57:103–112 16. Besl PJ, McKay ND (1992) A method for registration of 3-D shapes. IEEE Trans Pattern Anal Mach Intell 14:239–256. https://doi.org/10.1109/34.121791 17. Di Angelo L, Di Stefano P, Pane C (2018) An automatic method for pottery fragments analysis. Measurement 128:138–148 18. Gomes L, Bellon ORP, Silva L (2014) 3D reconstruction methods for digital preservation of cultural heritage: a survey. Pattern Recogn Lett 50:3–14 19. Di Angelo L, Di Stefano P, Morabito AE (2018) C Pane: measurement of constant radius geometric features in archaeological pottery. Measurement 124:138–146 20. Carr J, Beatson R, Cherrie H, Mitchel T, Fright W, Mccallum B, Evans T (2001) Reconstruction and representation of 3D objects with radial basis functions. In: Proceedings of the 28th annual conference on computer graphics and interactive techniques (SIGGRAPH), pp 67–76 21. Turk G, O’Brien J (2002) Modelling with implicit surfaces that interpolate. ACM Trans Graph 21(4):855–873. https://doi.org/10.1145/571647.571650 22. Dey TK, Sun J (2005) An adaptive MLS surface for reconstruction with guarantees. In: Proceedings of the third Eurographics symposium on Geometry processing, July 04–06 Vienna, Austria 23. Kolluri R (2008) Provably good moving least squares. ACM Trans Algorithms 4(2):1–25. https://doi.org/10.1145/1361192.1361195 24. Yang Z, Seo YH, Kima TW (2010) Adaptive triangular-mesh reconstruction by mean-curvature based refinement from point clouds using a moving parabolic approximation. Comput Aided Des 42(1):2–17. https://doi.org/10.1016/j.cad.2009.04.014 25. Amenta N, Choi S, Dey TK, Leekha N (2007) A simple algorithm for homoeomorphic surface reconstruction. Int J Comput Geometry Appl 12(1 & 2):125–141 26. Dey TK, Giesen J, Hudson J. Delaunay based shape reconstruction from large data. In: Proceeding of the IEEE symposium on parallel and large-data visualization and graphics 19–27 (2001). https://doi.org/10.1109/PVGS.2001.964399

1592

L. Di Angelo et al.

27. Cohen-Steiner D, Da F (2004) A greedy Delaunay-based surface reconstruction algorithm. Vis Comput 20(1):4–16. https://doi.org/10.1007/s00371-003-0217-z 28. Bernardini F, Mittleman J, Rushmeier H, Silva C, Taubin G (1999) The ball-pivoting algorithm for surface reconstruction. IEEE Trans Vis Comput Graph 5(4):349–359. https://doi.org/10. 1109/2945.817351 29. Huang J, Menq CH (2002) Combinatorial manifold mesh reconstruction and optimization from unorganized points with arbitrary topology. Comput Aided Des 34(2):149–165 30. Li X, Han CY, Wee WG (2009) On surface reconstruction: a priority driven approach. Comput Aided Des 41(9):626–640 31. Di Angelo L, Di Stefano P, Giaccari L (2011) A new mesh-growing algorithm for fast surface reconstruction. Comput Aided Des 43(6):639–650 32. Di Angelo L, Di Stefano P, Giaccari L (2013) A fast mesh-growing algorithm for manifold surface reconstruction. Comput Aided Des Appl 10(2):197–220 33. Pottmann H, Peternell M, Ravani B (1999) An introduction to line geometry with applications. Comput Aided Des 31:3–16 34. Cao Y, Mumford D (2002) Geometric structure estimation of axially symmetric pots from small fragments. Signal processing, pattern recognition, and applications, IASTED international conference 35. Di Angelo L, Di Stefano P, Morabito AE (2015) A robust method for axis identification. Precision Eng 39:194–203 36. Mara H, Sablatnig R (2006) Orientation of fragments of rotationally symmetrical 3d-shapes for archaeological documentation. In: Third international symposium on 3D data processing, visualization, and transmission, IEEE, pp 1064–1071 37. Halir R (1999) An automatic estimation of the axis of rotation of fragments of archaeological pottery: a multi-step model-based approach. In: Proceedings of the 7th international conference in Central Europe on computer graphics, visualization and interactive digital media (WSCG99) 38. Han D, Hahn HS (2014) Axis estimation and grouping of rotationally symmetric object segments. Pattern Recogn 47:296–312 39. Son K, Almeida EB, Cooper DB (2013) Axially symmetric 3D pots configuration system using axis of symmetry and break curve. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 257–264 40. Di Angelo L, Di Stefano P (2018) Axis estimation of thin-walled axially symmetric solids. Pattern Recogn Lett 106:47–52 41. Di Angelo L, Di Stefano P, Morabito AE (2014) Comparison of methods for axis detection of high-density acquired axially-symmetric surfaces. Int J Interact Des Manuf 8(3):199–200 42. Willis A, Orriols X, Cooper DB (2003) Accurately estimating Sherd 3D Surface Geometry with Application to Pot Reconstruction. Accurately Estimating Sherd 3D Surface Geometry with Application to Pot Reconstruction IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops, 1, art. no. 4624519 43. Hlavackova-Schindler K, Kampel M, Sablatnig R (2001) Fitting of a closed planar curve representing a profile of an archaeological fragment. In: Proceedings VAST 2001 virtual reality, archeology, and cultural heritage, pp 263–269 44. Kampel M, Mara H, Sablatnig R (2005) Robust 3D Reconstruction of Archaeological Pottery Base on Concentric regular Rills. In: Proceedings of the 6th international workshop on image analysis for multimedia interactive services, Montreau, Switzerland, April 13–15 2005, pp 14–20 45. Karasik A, Smilansky U (2008) 3D scanning technology as a standard tool for pottery analysis: practice and theory. J Archaeol Sci 35:1148–1168 46. Cohen F, Liu Z, Ezgi T (2013) Virtual reconstruction of archaeological vessels using expert priors and intrinsic differential geometry information. Comput Graph 37:41–53 47. Gilboa A, Tal A, Shimshini L, Kolomenkin M (2013) Computer-based, automatic recording and illustration of complex archaeological artifacts. J Archaeol Sci 1329–1339 48. Di Angelo L, Di Stefano P, Pane C (2017) Automatic dimensional characterization of pottery. J Cult Herit 26:118–128

53

3D Metrology for Ancient Pottery Classification and Reconstruction

1593

49. Di Angelo L, Di Stefano P, Morabito AE (2018) Secondary features segmentation from highdensity tessellated surfaces. Int J Interact Des Manuf 12(3):801–809 50. Solbrig T, Brunner D, Brunnett G (2006) Finding ornamentations on Bronze Age vessels to support the automated classification, EVA 2006 Berlin Conference Proceedings, pp 72–78 51. Karasik A, Smilansky U (2006) Computation of the capacity of pottery vessels based on drawn profiles. In: Mazar A (ed) Excavations at Tel Beth Shean 1989–1996, vol I. Appendix 1A to Chapter 12. (Jerusalem) 52. Senior LM, Birnie DP III (1995) Accurately estimating vessel volume from profile illustrations. Am Antiq 60(2):319–334 53. Pavić D, Kobbelt L (2008) High-resolution volumetric computation of offset surfaces with feature preservation. Comput Graphics Forum 27(2):165–174. Blackwell Publishing Ltd., Oxford, UK 54. Liu S, Wang CC (2010) Fast intersection-free offset surface generation from freeform models with triangular meshes. IEEE Trans Autom Sci Eng 8(2):347–360 55. Spelitz S (2017) Inner surface reconstruction of 3D scanned vessels. In: 3rd IMEKO international conference on metrology for archaeology and cultural heritage, IMEKO, pp 75–80 56. Soler F, Melero FJ, Luzón MV (2017) A complete 3D information system for cultural heritage documentation. J Cult Herit 23:49–57 57. Anichini F, Fabiani F, Gattiglia G, Gualandi ML (2012) A database for archaeological data recording and analysis. MapPapers 1en-II 21–38 58. Classical Art Research Centre. Beazley Archive Pottery Database (BAPD). http://www.beazley. ox.ac.uk/pottery/default.htm 59. Verga F, Fabbri F (2011) Un approccio globale, archeologico, archeometrico, informatico, alla schedatura della ceramica a vernice nera della Valle del Tevere. Archeologia e Calcolatori 22: 51–65 60. Smith NG, Karasik A, Narayanan T, Olson ES, Smilansky U, Levy TE (2014) The pottery informatics query database: a new method for mathematic and quantitative analyses of large regional ceramic datasets. J Archaeol Method Theory 21(1):212–250 61. Wilczek J, Monna F, Jébrane A, Chazal CL, Navarro N, Couette S, Smith CC (2018) Computerassisted orientation and drawing of archaeological pottery. J Comput Cult Herit (JOCCH) 11(4):22 62. Willis AR, Cooper DB (2008) Computational reconstruction of ancient artefacts: from ruins to relics (2008). IEEE Signal Process Mag 25(4):65–83 63. Grosman L, Smikt O, Smilansky U (2008) On the application of 3-D scanning technology for the documentation and typology of lithic artifacts. J Archaeol Sci 35(12):3101–3110 64. Biasotti S, Moscoso Thompson E, Spagnuolo M (2018) Experimental similarity assessment for a collection of fragmented artifacts, Eurographics workshop on 3D object retrieval, EG 3DOR 103–110 65. Marie I, Qasrawi H (2005) Virtual assembly of pottery using Moire surface profile measurements. J Archaeol Sci 32(10):1527–1633 66. Lu Y, Gardner H, Jin H, Liu N, Hawkins R, Farrington I. Interactive reconstruction of archaeological fragments in a collaborative environment. In: Proceedings 9th biennial conference Australian pattern recognition society on digital image (DICTA 2007), IEEE, pp 23–29 67. Liu J, Zhou M, Geng G, Xiamixiding R (2017) Broken pottery relic reassembling based on mixed feature vector. Boletin Tecnico/Tech Bull 55(5):1–9 68. Kampel M, Sablatnig R. 3D Puzzling of Archaeological Fragments. In: Proceedings of the 9th computer vision Winter Workshop, Piran, Slovenia, 4–6 February 2004, pp 31–40 69. Zheng SY, Huang RY, Wang Z, Li J (2014) Reassembling 3d thin fragments of unknown geometry, cultural heritage. ISPRS Ann Photogramm Remote Sens Spatial Inf Sci 2(5):393–399 70. Willis AR, Cooper DB (2004) Bayesian assembly of 3d axially symmetric shapes from fragments. In: Proceedings of the 2004 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, pp 82–89 71. Garcia Castaeda A, Brown B, Rusinkiewicz S, Funkhouser T, Weyrich T (2011) Global consistency in the automatic assembly of fragmented artefacts. In: Proceedings of the 12th international symposium on virtual reality, archaeology and cultural heritage VAST, Prato, Italy

1594

L. Di Angelo et al.

72. Orton C, Hughes M, Hughes M (2013) Pottery in archaeology. Cambridge University Press, Cambridge 73. Fish PR (1978) Consistency in archaeological measurement and classification: a pilot study. Am Antiq 43(1):86–89 74. Martinez Carrillo AL (2008) Computer Application in archaeological pottery: a review and new perspective, on the road to reconstructing the past. In: Proceeding of the 36th computer application and quantitative method in archaeology conference, Budapest, 2–6 April 2008, pp 397–401 75. Karasik A, Smilansky U (2011) Computerized morphological classification of ceramics. J Archaeol Sci 38(10):2644–2657 76. Sablatnig R, Menard C (1999) On estimating the position of fragments on rotational symmetric pottery, 3-D Digital Imaging and Modeling. In: Proceedings second international conference, IEEE, pp 455–462 77. Kampel M, Sablatnig R (2007) Rule based system for archaeological pottery classification. Pattern Recogn Lett 28:740–747 78. Porter D, Werner P, Utcke S (2005) Ancient ceramics: computer aided classification. Technical report M-338. University of Hamburg, Computer Science Department, KOGS 79. Gilboa A, Karasik A, Sharon I, Smilansky U (2004) Towards computerized typology and classification of ceramics. J Archaeol Sci 31:681–694 80. Adan-Bayewitz D, Karasik A, Smilansky U, Asaro F, Giauque RD, Lavidor R (2009) Differentiation of ceramic chemical element composition and vessel morphology at a pottery production centre in roman galilee. J Archaeol Sci 36:2517–2530 81. Koutsoudis A, Pavlidis G, Liami V, Tsiafakis D, Chamzas C (2010) 3d pottery content-based retrieval based on pose normalisation and segmentation. J Cult Herit 11:329–338 82. Hörr C, Brunner D, Brunnett G (2007) Feature extraction on axially symmetric pottery for hierarchical classification. ComputAided Des Appl 4:375–384 83. Roman-Rangel E, Jimenez-Badillo D, Marchand-Maillet S (2016) Classification and retrieval of archaeological potsherds using histograms of spherical orientations. J Comput Cult Herit (JOCCH) 9(3):17 84. Lucena M et al (2017) Classification of archaeological pottery profiles using modal analysis. Multimed Tools Appl 76(20):21565–21577 85. Karasik A, Smilansky U (2008) 3d scanning technology as a standard archaeological tool for pottery analysis: practice and theory. J Archaeol Sci 35:1148–1168 86. Maiza C, Gaildrat V (2005) Automatic classification of archaeological potsherds. In: The eight international conference on computer graphics and artificial intelligence, pp 11–12 87. Martínez-Carrillo AL, Ruiz-Rodríguez A, Lucena M, Fuertes JM (2009) A proposal of ceramic typology based on the image comparison of the profile. Comput Appl Archaeol 1e7 88. Zvietcovich F, Navarro L, Saldana J, Castillo LJ, Castaneda B (2016) A novel method for estimating the complete 3D shape of pottery with axial symmetry from single potsherds based on principal component analysis. Digital Appl Archaeol Cult Herit 3(2):42–54 89. Scopigno R, Cignoni P, Pietroni N, Callieri M, Dellepiane M (2017) Digital fabrication techniques for cultural heritage: a survey. Comput Graphics Forum 36(1):6–21 90. Echavarria KR, Samaroudi M (2018) Digital workflow for creating 3D puzzles to engage audiences in the interpretation of archaeological artefacts. In 16th EUROGRAPHICS Workshop on Graphics and Cultural Heritage (Workshop on Graphics and Cultural Heritage), Vienna, 12–15 November, Eurographics Association: 77–85

54

Mapping Stone Age Sites by Topographical Modelling: Problems and Possibilities Ole Grøn, Anton Hansson, Jessica Cook Hale, Caroline Phillips, Annabell Zander, Daniel Groß, and Bjo¨rn Nilsson

Contents 54.1 54.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Topographical Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.2.1 Historical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.2.2 Geological Modelling Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.2.3 Environmental Modelling Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1597 1598 1598 1603 1606

O. Grøn (*) Department of Geosciences and Natural Resource Management, University of Copenhagen, Copenhagen K, Denmark A. Hansson Department of Geology, Lund University, Lund, Sweden e-mail: [email protected] J. C. Hale Department of Anthropology, University of Georgia, Athens, GA, USA e-mail: [email protected] C. Phillips Anthropology, University of Auckland, Auckland, New Zealand e-mail: [email protected] A. Zander Department of Archaeology, University of York, York, UK e-mail: [email protected] D. Groß Centre for Baltic and Scandinavian Archaeology (ZBSA), Schleswig-Holstein State Museums Foundation, Schleswig, Germany e-mail: [email protected] B. Nilsson Department of Archaeology and Ancient History, LUX, Lund University, Lund, Sweden e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_54

1595

1596

O. Grøn et al.

54.3

Examples of Different Modelling Approaches and Tests/Estimations of Their Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.1 The Danish Fishing-Site Model as Used in Maritime Archaeology . . . . 54.3.2 Testing the Fishing-Site Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.3 The Southeastern US Outer Continental Shelf . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.4 Test of Different Modelling Methods: The Southeastern US Outer Continental Shelf . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.5 Locational Analysis Models in New Zealand . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.6 Evaluation of Locational Analysis Models Along the Waihou River, New Zealand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.7 Agent-Based Modelling (ABM) in the Southern Hebrides Mesolithic Project (SHMP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.3.8 Evaluating the Agent-Based Modelling (ABM) Employed in the Southern Hebrides Mesolithic Project (SHMP) . . . . . . . . . . . . . . . . . . 54.4 Examples of Complex Landscape and/or Cultural Features . . . . . . . . . . . . . . . . . . . . . 54.4.1 Hidden Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.4.2 Seasonally Variable Landscape Features: An Example . . . . . . . . . . . . . . . . . 54.4.3 Variation Over Time in an Area with Very Little Topographical Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.4.4 Lakes and Rivers: The Effect of Deep Lakes as Local Environmental Modifiers and the Effect of “Naled” Areas as Resource Magnets . . . . . . 54.4.5 Ideologically and Socially Based Variations in Landscape Behavior . . . 54.4.6 Cultural Small-Scale Variation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1607 1607 1609 1613 1614 1615 1616 1617 1619 1620 1620 1621 1623 1624 1628 1628 1631 1633

Abstract

Topographical/bathymetric archaeological modelling of potential Stone Age settlement zones has been criticized for operating in a world of its own and producing results with a poor relation to the real world. As landscape archaeologists who use modelling and an awareness of the importance of being able to detect Stone Age sites effectively in relation to GIS-based management and protection as well as research, we find it important to extend the debate about modelling. This includes its more advanced forms such as agent-based modelling (ABM) to a point where it interfaces in a convincing way with the available archaeological, ethnographic/social anthropological, landscape ecological, as well as geomorphological data and knowledge. To obtain this goal, it has been necessary not to base this review on the generally accepted presumptions inherent in archaeological topographical modelling of potential Stone Age settlement zones, but to approach this theme “from the outside” in a wide interdisciplinary perspective with a special focus on the highly relevant data from ethnography/social anthropology, landscape ecology, as well as geomorphology. This chapter aims to sum up the difficulties in topographical settlement zone modelling as well as its possibilities for making progress. It offers a review of the historical and theoretical background for modelling of potential Stone Age settlement zones in relation to the debate about environmental determinism in anthropology as well as the geomorphological and landscape ecological

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1597

backgrounds for topographical settlement zone modelling. Examples of modelling methods ranging from the rather basic to the quite complex, such as agentbased modelling (ABM), are used to give an idea of how different types of topographical modelling work. Tests/evaluations of their results are used to indicate how well they work. To elucidate the types of complexities, a number of examples are presented of behavior from recent small-scale cultures who demonstrate dynamic cultural variation.

54.1

Introduction

Topographical/bathymetric archaeological modelling of potential Stone Age settlement zones has been criticized for operating in a logical world of its own and producing results with a poor relation to the real world. As landscape archaeologists with a user role in relation to modelling and an awareness of the importance of being able to detect effectively Stone Age sites in relation to GIS-based management and protection as well as research, we find it important to extend the debate about modelling, including its more advanced forms such as agent-based modelling (ABM), to a point where it interfaces in a convincing way with the available archaeological, ethnographic/social anthropological, landscape ecological, as well as geomorphological data and knowledge. To reach this goal, it has been necessary not to base this method-review on the generally accepted presumptions inherent in archaeological topographical modelling of potential Stone Age settlement zones, but to approach this theme “from the outside” in a wide interdisciplinary and full historical perspective with a special focus on the highly relevant data from ethnography/social anthropology, landscape ecology, as well as geomorphology. This chapter aims to review and test/evaluate the different approaches to topographical modelling of Stone Age settlement zones to identify their strengths, weaknesses, and future potential and thus establish how useful this type of technology can be for practical archaeological prospection and research in the future. After a general introduction (Sect. 54.1), the historical and theoretical background for modelling of potential Stone Age settlement zones is outlined (Sect. 54.2), with its roots going back to the start of the quite intense environmental determinism debate in anthropology and other disciplines in the late nineteenth century. It is described how this leads to a virtual split in the 1960s between a modelling-based archaeology maintaining that technologically similar cultures will behave similar in similar landscapes and an ethnographic/social anthropological experience-based understanding maintaining that this is not the case due to historically based cultural/ideological differences. Section 54.2 also discusses the important geomorphological/geological and landscape ecological contexts that are not fully taken into account in their full complexity in topographical settlement zone modelling. Section 54.3 of this chapter provides four examples of topographical analysis and attempts to test/evaluate their results. The examples span from rather primitive applied modelling to mathematically very complex modelling (e.g., agent-

1598

O. Grøn et al.

based modelling – ABM) and include approaches based solely on archaeological data as well as approaches including historical and ethnographical/ethnoarchaeological information. This is to represent topographical settlement zone modelling through a number of practical examples of a differing character that together more or less expand this entire thematic field. The tests/evaluations of the different modelling approaches reveal that they are far from as reliable as one should have hoped. To elucidate the types of complexities one has to take into account when modelling the intricate interaction between human small-scale societies and nature, Sect. 54.4 presents a number of examples of landscape behavior from prehistoric or recent small-scale cultures that deviates from what one would typically assume. Landscapes and their availability of resources can display an annual variation or changes over longer time intervals that can be important for their users but difficult to distinguish today because the prehistoric environments and their dynamics are difficult or even impossible to reconstruct in sufficient detail. Small-scale environmental features such as deep lakes, the presence of “naled” areas, and local wind conditions in valleys are important for where the good settlement zones will be. In addition, small-scale cultures can display significantly different landscape behavior due to ideology/religious ideas, and even internally their subgroups display a strong identity-driven tendency to differentiate themselves from the other subgroups. The final discussion (Sect. 54.4.5) sums up the difficulties in topographical settlement zone modelling as well as its possibilities for making progress. From a modelling point of view, it is understandable that one wants to simplify modelling by focussing on theory and information that is relatively easy to handle and to exclude information with a structure that is difficult to handle mathematically. However, that is a dangerous course when it comes to the practical value of the results. Therefore, it is important to “reality check” the archaeological modelling methods and underlying assumptions as they develop. Critical for the further development of archaeological topographical settlement zone modelling is that it develops an ability to distinguish and incorporate landscape behavioral differences between cultures and even between their subgroups as well as to cope to a reasonable degree with dynamic small-scale environmental variation.

54.2

Topographical Modelling

54.2.1 Historical Background Modelling the potential positions of Stone Age sites in landscapes on the basis of landscape topography (¼ bathymetry under water) has become a widely applied method in cultural heritage management as well as in research. The basic idea is that some types of topographical settings are more attractive than others and therefore on a general basis will attract a significant part of the settlements of all cultures operating in an area over time. Thus, focusing surveys on such features should

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1599

provide access to a sample of sites that can represent in some detail the cultural development in the area. The term “predictive modelling” is often used for such approaches. This term, however, also includes many other types of prediction such as prediction based on soil types, proximity to water bodies, geological data, or other types of landscape data [1–3:18]. Here, we focus on the types of modelling and interpretation that have a strong focus on topography/bathymetry, which is most of the applied methods. An important motivating driver behind the development and application of topographical modelling is that it can save time and money if it can be developed to an efficient tool [2, 3:13]. The more data one includes in the prediction, the more time-consuming and expensive the method will become. The aim of this chapter is to investigate in a critical way how much detail is needed to obtain satisfying results with topographical/bathymetric modelling – also in a critical perspective – to review the historical and theoretical basis for topographical modelling of the positions of Stone Age sites in general. Historically, topographical modelling goes back to field archaeological experience concerning where Stone Age sites were observed in the landscape. Quite typical for this type of approach is, for instance, the description of the Danish Maglemose sites as “inland stations positioned on small islands or peninsulas in lake basins quite some distance from the sea” [4:47,70]. This is a reasonably concise description of a large body of the sites found at that time, providing the field archaeologist with a “model” that allows him to find more rather visible – or exposed – Maglemose sites [2]. Such a model may, however, not lead him to sites in positions that tend to be covered by significant later earth masses, such as sites adjacent to hills with a significant sediment flow [5], and therefore not provide a true picture of the site distribution and settlement pattern. In consequence, this easily leads to a strategic focussing on already known skewed positioning patterns and the suppression of less easily observable ones. The discussion of cultural ecology and environmental deterministic man-land relationship in hunter-gatherer societies started by Ratzel’s Anthropogeografie (1909) [6] of which vol. 1 was published in 1882 and was re-vitalized by Steward’s culture ecological work in the 1930s. His idea that hunter-gatherer “culture, environment, and subsistence adaptation” are inextricably linked and determine where their sites are located formed the basis for attempts to predict their settlement locations as a general function of the basis of central environmental/landscape features such as the resources in combination with the cultural technological level [1, 7–12]. A critical point here, however, is the precise reconstruction of the resources and their distribution in the contemporary prehistoric landscapes. As Steward puts it [10:42]: Cultural diffusion, of course, always operates, but in view of the seeming importance of ecological adaptations its role in explaining culture has been greatly overestimated. The extent to which the large variety of world cultures can be systematized in categories of types and explained through cross-cultural regularities of developmental process is purely an empirical matter. Hunches arising out of comparative studies suggest that there are many regularities which can be formulated in terms of similar levels and similar adaptations.

1600

O. Grøn et al.

Hunter-gathering cultures are seen as especially well-suited for ecologically based prediction because they more directly than other culture types are depending on their environment. Therefore, they have been more in focus than other culture types in prediction studies [1]. The consequently environmental deterministic line of thinking starting with Steward’s work has from the beginning been under fire from the opposite position, that human culture – independent of its technology – plays a role for how the man-environment relation plays out. Researchers such as Boas [13], Forde [14:460–466], Herskovits [15:153–166], Kroeber [16:205–217], Wallis [17], and Wissler [18] are significant early opponents. Typical for this opposition is Wallis’ statement [17]: In explaining everything the environment explains nothing.... If we wish to predict what a people will do when they move into a new environment, it is more important to know the people than to know the place – or better, one must know both.

Interesting is Herskovits’ argument that if the environment is an important determinant of culture, similar environments should be inhabited by similar cultures. He demonstrates through examples that similar cultures existed in different environments. He also demonstrates that different cultures existed in similar environments – such as the Inuit and some Arctic Siberian tribes that lived in similar environments but interacted with them in quite different ways [15:158]: Here in the difficult circumpolar habitat then we have two quite different ways of life. . . The adaptation of both peoples is equally successful inasmuch as the only test of success in adaptation is survival.

An increased focus on ecology and human man-land relations developed in archaeology in the early 1960s. This development was among other things related to a strong positivistic – “processual” – development reflected in the appearance of “new” archaeology and a related positivistic focus on ethnoarchaeology seen as a “live laboratory” for archaeology [8, 19, 20:29–45]. As the man-land relations were understood within the frames of rather primitive and generalizing natural scientific models, to a high degree based on animal behavior and regarding human societies as behaviorally uniform [19, 21–23], this led to a decreasing focus on the variability of the individual cultures which facilitated a strengthening of the environmental deterministic positions. Spaulding’s attempt to create a basis for detailed statistical/spatial interpretation of archaeological cultures (“spatial archaeology”) by focusing on “the dimensions of archaeology” promoted statistical analysis of cultural features in mainly two- or three-dimensional space [24]. At the same time, however, this approach further legitimated an archaeological negligence of intra-cultural variations in spatial behavior in the landscape both between synchronous prehistoric cultures with no geographical overlap and geographically overlapping diachronous prehistoric cultures. The focus was on the development of general/generalizing culture models that did not give much attention to the individual cultures’ differences in landscape behavior and exploitation [7, 25, 26:1–95], [9:3–10], [3:14].

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1601

Parallel to this development in processual archaeology and its related ethnoarchaeology, the development in social anthropology took a different general direction distancing itself from any strict environmental determinism that could allow identification of potential settlement areas on the basis of global principles [27, 28]. Here as well as in more social-anthropologically orientated ethnoarchaeology, it is underlined that even though ecology always puts some restrictions on a culture’s behavior in the landscape, cultural traditions independent of the actual environment also contain significant economic, social, and spiritual features which influence the settlement patterns [29–32], or as Barth puts it [27:13]: . . .., we need a viewpoint that does not confuse the effects of ecologic circumstances on behaviour with those of cultural tradition, but which makes it possible to separate these factors and investigate the non-ecological cultural and social components creating diversity.

The “historical determinism” of the structuralist part of modern social anthropology, which focuses on the study of the historical development of structural features within the social-spiritual spheres of societies, leaves only little space for an environmental determinism [33:1–28, 229–239] and represents, thus, an understanding directly opposed to that of the New Archaeology which, for instance, generally ignores the effect of ideology and spiritual phenomena [19]: This reasoning is functionally linked to a methodology that limits the archeologist to generalizing about the “facts” he uncovers. Since preservation is always imperfect, inferences from the facts of material culture to statements about the non-material culture move us away from the primary data and thus diminish the reliability of our statements.

This isolation of archaeology and archaeology-related ethnoarchaeology from the complex and culturally variable behavioral systems dealt with in social anthropology both ignoring the general environmental determinism debate in anthropology before the 1970s and in social anthropology after that [33, 34:186] facilitated a further narrowing down of the landscape factors taken into account in the Geographical Information Systems (GIS) now handling large amounts of data. The term “environmental determinism” is in this context generally avoided possibly due to its earlier quite sensitive role supporting racism and imperialism [34:169–195]. According to Verhagen [3:13]: Predictive modelling departs from the assumption that the location of archaeological remains in the landscape is not random, but is related to certain characteristics of the natural environment. The precise nature of these relations depends very much on the landscape characteristics involved, and the use that prehistoric people may have had for these characteristics; in short, it is assumed that certain portions of the landscape were more attractive for human activity than others. If, for example, a society primarily relies on agricultural production, it is reasonable to assume that the actual choice of settlement location is, among others, determined by the availability of suitable land for agriculture.

1602

O. Grøn et al.

The lack of inclusion of “socio-cultural” factors is, however, recognized as a problem [3:20]: Socio-cultural factors are virtually absent in predictive modelling studies up to now, and methods for using them in a predictive modelling context will have to be developed more or less from scratch.

In practice, this often boils down to modelling on the basis of topography, since this is generally the only type of high-resolution landscape data easily available, often supported by other rather generalized and/or roughly reconstructed surface quality data such as soil types, vegetation, slope, elevation, water, and reconstructed micro climate [35–40]. During its post-processual period, theoretical scientific archaeology did not care much about practical surveys and rescue archaeology, whereas practically orientated archaeology had only little interest in the more theoretical aspects of the discipline [39:18–19, 41]. As an empirical tradition has lately started gaining land, and direct detection methods for submerged Stone Age sites are under development [42, 43], criticism of the basic problems in the application of the – now GIS-based – topographical modelling has been intensifying [43–48]. Parallel to that, attempts are being made to solve the problem of social-cultural variation in the modelling through distinction of areas and chronological zones with differences in landscape behavior [49, 50] as well as through ABM – agent-based modelling [51, 52]. One central question for ABM with regard to prediction of potential settlement locations in the landscape is whether synthetic modelling on the basis of the often non-representative cultural data available can simulate in sufficient detail the complicated systems of dynamic small-scale cultural variation that can be observed in recent hunter-gatherers as well as related small-scale societies [53:55–64, [54]:75, 55–57]. Since we still struggle with understanding the logics/rationales of such live culture groups, the formulation of rationales/logics for synthetic agents is a weak point [58, 59]. Decisions can in some situations have “turning-point” consequences, which are difficult to model, for the “acting” groups as well as for their immediately uninvolved neighboring groups. Another central question is whether it will be possible to reconstruct the dynamics of the environment in sufficient historical detail to mimic a realistic basis for specific important prehistoric events in relation to the specific cultures exploiting it and living in it. It is obviously important to bridge the gap between the disciplines that are relevant for the study of human small-scale cultures in prehistory. That information from social anthropology/ethnography is regarded as “difficult to handle” or “unproductive” [34:186] does not legitimate an exclusion of it from archaeology. With the broad application topographical modelling and the related types or predictive modelling have gained in archaeological management, it is important to clarify what their qualities and restrictions are. It is important to secure that the assumptions applied in any kind of modelling are in accordance with the well-established knowledge social anthropology has accumulated about the types of societies to be modelled to avoid the type of problem pointed out by Wandsleeben and Verhart [40]

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1603

in their paper “Geographical Information Systems. Methodological Progress and Theoretical Decline?” From archaeology’s side, one can in theoretical discussions accept or reject the cultural small-scale diversity (see Sect. 54.4.6), which appears in the social anthropological/ethnographic studies of live cultures as a general feature of the prehistoric societies as well. When the purpose, however, is in practice and as precisely as possible to model the positions of concrete prehistoric sites in the landscape, it is important to operate on the basis of realistic information about the landscape behavior of the prehistoric human populations in focus. On this point, social anthropology is through its body of empirical data considerably better informed than archaeology, as well as landscape ecology is where one should look for information about the spatial behavior in landscapes of plant and animal societies. We have therefore found it necessary to accept as a fact and a precondition for such modelling the picture provided by social anthropology of small rather autonomous groups with a high degree of dynamic cultural small-scale variation.

54.2.2 Geological Modelling Conditions Manipulation of a topography or bathymetry dataset by shifting the water level relative to the present shoreline is an established method for reconstructing past landscapes, which forms the basis for modelling of archaeological sites. However, in order to perform accurate reconstructions of past landscape features, three main factors – bathymetry/topography, sea level changes/land uplift, and sediment infill – should be taken into account. The accuracy of these factors will reflect the accuracy of the reconstructed landscape features.

Bathymetry/Topography A bathymetry or topography data set is the foundation for any landscape reconstruction. The resolution of such a data set can vary for various reasons, such as the size of the area, the technique with which the data was gathered, and legal restrictions. Today, it is possible to acquire detailed elevation data sets over large areas. On land, the LiDAR (light detection and ranging) technique, is based on the emission and return of laser pulses from a sensor on an aircraft [60]. The LiDAR technique also has the advantage that it can remove vegetation from the elevation data, leaving only the true ground measurements. Below water, echo sounding is used to gather depth data. An echo sounder measures the time interval between emission and return of sound pulses from a ship. The measured time is then used to calculate the water depth. The water depth can be obtained either by a single sound pulse (single beam echo sounding) or by an array of multiple pulses (multibeam echo sounding). Contrary to single beam techniques, a multibeam echo sounder emits a simultaneous array (normally 120 – 150 ) of sound pulses, which enables the production of extremely high-resolution bathymetric data sets [61]. Multibeam echo sounding provides data sets with a

1604

O. Grøn et al.

horizontal error of a small percentage of the water depth and a vertical error of less than 1% of the water depth. The data set can then be used to produce a Digital Elevation Model (DEM) of an area of interest. The final resolution can vary, from less than 1 m per pixel in small sites such as Haväng, Sweden [62], to tens of meters over larger areas, such as the Hanö Bay in Sweden [63], where military interests limited the resolution available to 100 m per pixel in certain areas.

Sea Level Changes/Land Uplift Accurate shoreline displacement is a necessity for successful reconstructions in coastal settings. There are several methods for reconstructing past water level fluctuations, above and below the present shoreline. During the last glacial maximum, the global sea level was around 130 m below present [64]. Globally, the sea level stabilized around 6000 years ago, and therefore, sea level fluctuations need to be considered in all coastal Stone Age settings in order to correctly understand the past landscape features. In formerly glaciated areas, isostatic land uplift plays a major role in the landscape formation. The interplay between the global water level fluctuations and the local land uplift results in, sometimes very rapid, changes in the local shoreline displacement. Rapid water level changes, such as the Ancylus Lake transgression in the Baltic Sea around 10,500 years ago where the water level rose with about 4 cm/year during 500 years [65], stress the importance of a robust chronology in order to avoid erroneous placement of an archaeological site relative to the shoreline. Above the present shoreline, two commonly used methods are the study of raised beaches and the isolation of basins. The isolation method reconstructs the past shoreline displacement by dating the transition from fresh to saline water (or from saline to fresh water) by studying the diatom flora in the basin. In combination with a chronology of the lake stratigraphy and the elevation of the basin threshold, the time and elevation when the sea stood at the same level as the threshold can be determined. This method has been used across the globe, e.g., in Greenland [66], Russia [67], and Sweden [68]. Raised beaches record the highest past shoreline in an area, and by dating material embedded in the raised beach, the former water level can be determined. However, the elevation of a raised beach most often indicates the water level during heavy storms, and the mean past water level is likely to have been about 1–2 m below the raised beach [69]. The study of raised beaches has been used, for example, in Finland [70] and on Iceland [71]. Reconstructions of water level fluctuations below the present shoreline are often less accurate than above the present water level. Submerged tree stumps have been used since the middle of the nineteenth century [72–74], although some of these findings were not found in situ and therefore no accurate shoreline levels could be determined. Most often, a submerged tree stump provides information saying that the past water level was lower than the position of the stump, but how much lower is not possible to determine. The Haväng area in southeastern Skåne constitutes an exception to this, where large quantities of submerged rooted tree stumps have been sampled and 14C-dated [62]. Careful surveys of the seafloor by divers determined the

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1605

elevation of the deepest rooted stump, which meant that the reconstructed water depth could be determined with an error of only a few meters. Other features below the present water level that can determine past water levels, more or less accurate, are archaeological sites and their associated findings or organic sediments from past river mouths and lakes.

Sediment Infill In order to obtain accurate landscape reconstructions, younger sediments should be removed so that the reconstruction visualizes the correct past landscape features. To perform this so-called strip back of younger sediments, sediment sequences from the area of interest need to be obtained, and a chronology of the sediment deposition needs to be determined. Depending on the size and nature of the reconstructed area, the problem with sediment infill can be approached differently. On a relatively small scale and in settings where the sediment deposition is believed to be uniform across the area of interest, the back stripping can be performed based on one or a few data points. In larger areas, the sediment deposition is most likely not uniform, which complicates the back-stripping process. One example is the Blekinge archipelago in southeastern Sweden, where the water level was situated about 20 m and 4 m below the present shoreline during the Yoldia Sea and Initial Littorina Sea Stages, respectively [63]. The area consists of exposed areas where little or no deposition has taken place, as evident by the findings of tree stumps dating to about 11,000– 10,500 cal BP [63, 75], and well-protected areas where several meters of deposition have occurred since the early Holocene [76]. The shift from one depositional environment to the other is depending on the local topography and bathymetry, and there is no simple overall solution. In this case, the solution was to make no correction for the infill of younger sediments [63]. One example of a successful strip back of younger sediments can be found in the Tolkuse-Rannametsa area in southwestern Estonia [77]. The modern area consists of a near-shore peatland where a few Stone Age sites have been found. In order to strip back the sediments, the peatland was thoroughly investigated ground-penetrating radar and several corings along transects in order to establish a stratigraphy and chronology of the peatland. Based on the initial results, the DEM was manipulated in order to correctly reconstruct the landscape features during the early Holocene. The results revealed rapidly shifting coastal landscape during the early Holocene, culminating in a coastal lagoon around 7000 cal BP. In the reconstructed landscape, the Stone Age sites were placed at the river mouth of a river entering the lagoon. Together, the factors discussed above form the foundation for accurate reconstructions of past landscape features. However, the importance of one more factor, a robust chronology, cannot be understated in the landscape reconstruction process. This is especially true for the shoreline displacement, where rapid water level shifts can mean that a bad age control can misplace the reconstructed shoreline with several meters and potentially place well-dated archaeological sites seemingly under water.

1606

O. Grøn et al.

54.2.3 Environmental Modelling Conditions Important for the choice of settlement areas in the landscape is, apart from the topography/bathymetry to a high degree, the spatial distribution and population dynamics of the individual animal and vegetation species of importance for settlement. In widely applied approaches such as site-catchment analysis of the resources available from a site, it is generally assumed rather naively – and often without stating it clearly in the text – that these have a more or less even and stable distribution in the landscape and that the landscape apart from that can be modelled as a more or less isotropic surface [21, 78]. Resources representing significant and economically important deviations from such a pattern are, for instance, large-scale migrating mammals (e.g., reindeer, elk), birds, and fish [79, 80:12–16, 81, 82]. But also the vegetation displays under natural and uncontrolled conditions a high degree of dynamics with regular appearance of natural wildfires, which, within some areas as often as every tenth year, will radically change the vegetation and thus also the whole resource pattern [29, 83]. Landscape ecology in its struggle to model and thus to be able to predict the small and large-scale variability in the resources in the landscape in the mid-1990s generally accepted that a new level of complex mathematical modelling had to be applied [84]: A marked change has occurred recently within the science of ecology. Previously, ecological processes commonly were assumed to proceed within homogeneous environments, and usually within populations of randomly distributed individuals. Recently it has been widely recognized that environments are not homogeneous, and organisms are usually clumped into patchy populations, and that this heterogeneity has significant effects on ecological processes.

This development was accompanied by a change of focus from landscapes consisting of complementary habitat/biotope mosaics separated by softer or sharper borders (ecotones) to landscapes consisting of dynamic and to a high degree overlapping population patches. The level of mathematical complexity in the description/modelling of the populations of different species as well as of their dynamics/interaction will in most cases exceed what one can hope realistically to apply to prehistoric landscape situations [84, 85], where alone a sufficiently precise dating of the different parts of spatiotemporal landscape developments will be a problem [45, 86]. If later research results pointing to that the prehistoric hunter-gatherers were actively manipulating their plant resources with systematic and controlled burning possibly back to 50–100 ky or possibly even further, are correct, this makes modelling even more complicated [45, 87–90]. Direct manipulation of the distribution patterns of animal populations as observed with the Evenk and recognized as “what one would do in such a situation” by a Sami informant [30] also points to that such complicated strategies are of some age.

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1607

At least in the case of the Evenk, but probably generally, hunter-gatherers are extremely well-informed about the resources in the landscapes they inhabit as well as their dynamics, so they to some degree are able to extrapolate and thus predict the different animal and plant species’ variation over time as well as their spatial behavior and interaction. This allows them to adapt their settlement pattern to well-founded hunting and gathering strategies taking into account dynamic details in the total resource situation [91:41–49, 92, 93]. This is the type of strategic thinking that forms the basis of the choice of locations to settle.

54.3

Examples of Different Modelling Approaches and Tests/ Estimations of Their Efficiency

54.3.1 The Danish Fishing-Site Model as Used in Maritime Archaeology The Danish fishing-site model is based on an old assumption in Danish archaeology that the central coastal Late Mesolithic sites, often with significant shell middens, were located adjacent to optimal fishing areas, typically at the mouth of streams meeting the sea, at the mouths and narrow parts of fjords, as well as on small islands and peninsulas near sloping sea floor [35]. Since submerged Stone Age sites in 1983 became protected by law, a wish arose in the Danish Cultural Agency of the Ministry of Culture to be able to model the positions of the submerged Stone Age sites on the basis of such topographical indicators. In 1993, the topographical/bathymetric “fishing-site model,” a somewhat simplified predictive approach to the observations and data at hand, was promoted at the agency’s own symposium: “Man and sea in the Mesolithic. Coastal settlement above and below present sea level” [35, 94] (Fig. 54.1). The model has been published as a series of sketches of generalized topographical coastal situations with the assumed Late Mesolithic settlement positions marked (Fig. 54.1). The administrative practice that developed was that funding for rescue archaeological survey and excavation, which the Danish Cultural Agency controlled, could be given only to the types of topographical situations sketched out in the model, whereas rescue investigations of other types of situations could not be funded. The administrative main focus became the model’s situation A, the mouths of inlets and fjords (Fig. 54.1), possibly because the other situations in practice were more difficult to distinguish. More than other maritime archaeological projects, the survey before the construction of the Big Belt Bridge helped profiling the method and the agency on the basis of Denmark’s Avant Garde role in Stone Age archaeology under water including its early development of a national legislation for the management of this type of cultural heritage [95,96]. On this basis, the Danish fishing-site model gained international recognition and has inspired similar or modified approaches internationally for management as well as research-based investigations [97,98,99].

1608

O. Grøn et al.

Fig. 54.1 Generalized rules for the topographic location of (Late) Mesolithic coastal settlements in South Scandinavia. (a) ¼ along a narrow inlet connecting large water surfaces and with considerable hinterland on both sides. Here, the most potential site locations are immediately beside the narrowest spot. (b) ¼ along a narrow inlet between a small island and the mainland. Here, the most potential site location would be on the mainland side. (c and d) ¼ at the tip of a headland. The probability of finding settlement remains is greatest if the headland juts into sheltered water without strong waves. (e and f) ¼ at the mouth of a larger stream or river. Here, the most potential site location is on relatively flat land (Figure and Text: [35])

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1609

54.3.2 Testing the Fishing-Site Model Fischer’s Tests of the Danish Fishing-Site Model A test of the Danish fishing-site model for prediction of the positions of Late Mesolithic coastal Stone Age settlements was carried by the Danish Cultural Agency in 1985–1986 in the Danish Småland Bight southwest of Zealand [100]. One important element for the development of the model was an observation from extensive survey in the shallow areas of the morphologically highly complex Karrebæk Fjord system led by Axel Degn Johansson, that the Late Mesolithic sites seemed to appear at locations, which are today optimal for pond nets or fish traps [101,102]. Another important element was the results of a survey in Roskilde Fjord carried out by Fischer and Sørensen in the autumn of 1982, which documented a massive appearance of submerged Kongemose and Ertebølle sites in certain characteristic topographical/bathymetric positions. One should, however, take into account that the choice of survey locations during this survey was conducted by a topographical model more or less identical to the resulting one [103]. A third important element was the survey in Øresund between Sweden and Denmark carried out by Lars Larsson [104]. A fourth important element was the general notion taught at least at Aarhus University by Søren H. Andersen that the central Ertebølle sites with shell middens generally were placed in positions adjacent to the best fishing areas – the mouths of fjords and inlets. A “tautological” approach to survey for submerged Stone Age sites was normal in the early Danish maritime Stone Age archaeology because the focus was on registering as many submerged Stone Age sites as possible to document the dimensions of the cultural heritage problem that had to be managed/solved [105:106–121]. The 1985–1986 “test” of the fishing-site model’s efficiency did also target topographical/ bathymetric situations known to have a high probability of the presence of Mesolithic settlement. Out of the 27 test locations chosen, 17 were in advance categorized as “high probability” locations, nine as “reasonably high probability” sites, and only one as a “low probability” location. Twenty-six of these locations, including the “low probability” one, produced Mesolithic settlement. Site 20, which was originally categorized as a “high probability” location but did not produce any evidence of submerged settlement, was subsequently re-categorized as a “low probability” site. On this basis, the method’s predictions were calculated to be 96% correct [100]. A second test of the Danish fishing-site model was the maritime archaeological survey for submerged Stone Age sites in advance of the construction of the Big Belt Bridge in 1986–1998. The published statistical analysis of the registration of Stone Age settlement in the topographically/bathymetrically selected test locations is here, however, very unclear. A table (Table 1 – [95]) shows the “average number of pieces of worked flint per hour of diving” and the “number of hours of diving” for the site categories “very suitable,” “fairly suitable,” “partly suitable,” and “unsuitable” without specifying the results for the individual sites [95]. One can deduct from Table 1 in Fischer’s text that 10.8 h of diving on the very suitable locations produced 444 pieces of worked flint. According to Fig. 54.2 (¼ Fig. 54.6 in [95]), none of these were related to the tracé. It is unclear from the text whether a number of sites of

1610

O. Grøn et al.

Fig. 54.2 The distribution of sites explored by diving which were assessed in advance using the fishing-site model. Suitability for coastal settlement: (a) very suitable. (b) fairly suitable. (c) partly suitable. (d) unsuitable [95]. The bridge tracé is limited by black broken lines

this category along the eastern shore of the belt shown in Fig. 54.2 are included in the survey and thus are represented in this figure. Since sites of this category are absent from the tracé, this seems to be the case. An explanation of this point is lacking. Fifty-five hours of diving on the partly suitable locations produced 330 pieces of worked flint. According to Fig. 54.3, there are three of these in the tracé near the belt’s eastern shore and one in Nyborg Bay (the Lindholm site) on its western shore, which was actually known in advance of the survey. Furthermore, there are 13 sites

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1611

Fig. 54.3 Distribution of known sites for the Early Holocene along the coastal plains of Georgia and Florida. The coastline position is that which is calculated for 11,000 cal BP for the Gulf of Mexico following Joy [116], and paleochannels were projected using bathymetric data and the ArcMAp 10.3 hydrological toolset. A subset of sites (N ¼ 136) were selected from the wider dataset (N ¼ 500) and then classified by aquifer conditions (confined versus unconfined) and proximity to the nearest coastline (the Gulf of Mexico versus the Georgia Bight). The histogram at the bottom demonstrates a significantly higher number of sites in the zone of unconfined aquifer closest to the

1612

O. Grøn et al.

of this category outside the tracé, which possibly, possibly not, contribute to this figure. 19.4 h of diving on the partly suitable locations produced altogether two pieces of worked flint. According to Fig. 54.2, there are seven locations of this category in the tracé. Furthermore, 14.2 h of diving on the unsuitable locations produced a total of one piece of worked flint. According to Fig. 54.2, there are eight locations of this category in the tracé [95]. One might get the impression that a number of productive sites outside the tracé have been pulled in to improve the statistics. A fact is that the seafloor in the main part of the Big Belt by the strong current had been eroded down to hard moraine displaying basins truncated by the current as well as up to 4-m-high reefs of large stones standing as the remains of softer moraine structures washed away. Because of this, the topographically/bathymetrically selected locations in the main part of the tracé, which – to shorten the bridge length – were placed in a bottleneck position in the belt, yielded very little in situ settlement material at all [106] (Fig. 54.1). The unpleasant conclusion of a thorough statistical analysis would have been that the application of the model to a significantly topographically/bathymetrically modified seafloor had not been very successful. This, however, is not obvious from the data presented [95].

Grøn’s Test of the Danish Fishing-Site Model A test of the fishing-site model was carried out by Grøn in 2013–2014 on the basis of the survey results obtained by application of the model to maritime archaeological rescue archaeology on the basis of the then available reports from the Viking Ship Museum in Roskilde and from Moesgaard Museum in Aarhus [29, 107]. The total surveyed sea area was found to amount to 317 square kilometers and to have yielded 15 hitherto unknown Stone Age settlements. To compare this with the situation on land, eight randomly chosen municipals with an “average” level of registration of their cultural heritage sites were found to cover a total area of 2193 square kilometers and have registered 3250 Stone Age settlements. Two intensely surveyed areas, the Aamose basin, and the Vedbæk area, both in Zealand were also included. They covered 34 square kilometers and had registered 291 Stone Age settlements. The number of settlements found in the surveyed marine areas was thus 0.05 per square kilometer in contrast to an average on land of 1.49 settlements per square kilometer and for the intensively surveyed areas of 8.61 settlements per square kilometer. One can say that the surveyed marine areas produced 0.6% of ä Fig. 54.3 (continued) Gulf of Mexico, despite the fact that multiple paleoclimate assessments for the Early Holocene have demonstrated that the coastal plain of northern Florida experienced periods of intense drought and aquifer drawdown during this period [108, 134]. At the same time, fluvial systems on the Georgia coastal plain in areas of confined aquifer closer to the Georgia Bight appear to have carried 1.5–4 times greater bankfull discharge than at present [135], making this region much more favorable, hydrologically speaking. Even without multivariate statistical analysis, it is clear that site distributions were not linearly related to environmental conditions, calling into question resource-driven occupation patterns by Early Holocene foraging groups [132]

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1613

what one should have expected to find on land in a similar area by intense survey [29, 107]. Of course, the change over time in sea-covered and not-sea-covered areas in Denmark will make such a figure a bit difficult to take at its face value. However, it will also be difficult to argue that the model has produced more than a couple of percent of the sites present in the surveyed areas. Thus, there seems to be a serious problem locating and protecting/managing submerged Stone Age settlements in the hypothetical settlement zones indicated by the model and/or in the assumptions about the location of Mesolithic settlements it is based on.

54.3.3 The Southeastern US Outer Continental Shelf Submerged prehistoric archaeology has focused intensively on the southeastern United States, particularly Florida and the northern Gulf of Mexico coastline, because the area already demonstrates some of the oldest sites in the Western Hemisphere [108–110] along with some of the densest occupations in North America as early as the terminal Pleistocene [111]. Critical questions of interest to global archaeology can be asked and answered here, including time depth of coastal foraging, human responses to dramatic coastline and ecological changes, and diversity in human foraging patterns, from the terminal Pleistocene to the opening of the Late Holocene, when coastlines stabilized here [112, 113]. Thousands of square kilometers of formerly subaerial coastal plain lie submerged today; much of it within the Gulf of Mexico along the western continental shelf of Florida and the South Atlantic Bight along the Georgia coastline [114–116]. Multiple examples exist of attempts to develop systematic approaches to detect high probability areas for submerged prehistoric sites since the 1970s, when the National Park Service funded the first of multiple studies aimed at paleolandscape reconstruction and identification of high probability zones for prehistoric archaeological sites [117–119]. However, offshore research in this region remains uneven. Along the South Atlantic Bight, only a few deflated sites have been documented [115]. Along the Florida Atlantic coastline, one submerged site, Douglass Beach, was documented around remnants of the 1715 Plate Fleet Wreck in the 1960s and 1970s [120, 121]. In the western Gulf of Mexico, Coastal Environments, Incorporated, developed a comprehensive landscape model that synthesized geological, geomorphological, and archaeological data, assessed areas for preservation potentials, and then developed a sedimentologically based model for detecting archaeological sediments using coring [117, 118]. From these baseline studies, they then launched geophysical surveys along Sabine Pass, Texas, followed by coring that yielded archaeological materials that are deeply buried [119, 122]. Later in the 1980s, the Minerals Management Service (MMS), Department of the Interior of the United States (now Bureau of Ocean Energy Management [BOEM]) funded work at McFaddin Beach on the north Texas coastline, where potential sources for archaeological materials remain highly enigmatic [123, 124]. Apalachee Bay, on the western coast of Florida, is the best understood after 30 years of research. Multiple sites and site types

1614

O. Grøn et al.

documented range in age from terminal Pleistocene to Middle to even Late Holocene [125–131]. Research programs have included varying degrees of use of predictive modelling to aid detection and recovery of archaeological materials in this region with varying degrees of success. Work along the South Atlantic Bight was guided by physiographic parameters (primarily paleochannels and rocky outcrops). Work at Sabine Pass was predicated on onshore archaeological depositional trends and attendant landforms [117–119]. Testing at Douglass Beach, Florida, involved no predictive modelling and has yet to be articulated with the paleolandscape. And work in Apalachee Bay closely followed regional archaeological and geohydrological trends onshore [126–128]. Findings in Apalachee Bay demonstrated the utility of the regional model used there, but these parameters are not scalable elsewhere. These mixed results suggest the need to incorporate additional high probability variables and call into question the utility of general landscape-based models.

54.3.4 Test of Different Modelling Methods: The Southeastern US Outer Continental Shelf In an effort to bridge the gap between overly general models such as those used along the Georgia coast, with highly specific models used to greater success in Apalachee Bay, one co-author, Cook Hale, worked to operationalize landscape variables within a geographic information system (GIS) and test for which variables correlated best with sites located on the coastal plain of Georgia and northern Florida [130–132]. Higher densities of Clovis points (terminal Pleistocene) correlate meaningfully to certain geological, geomorphological, and ecological parameters in the southeastern United States [133], but tests of terminal Pleistocene site locations on the coastal plain of Georgia suggested spatial variation in landscape indicators [130]. Wider study was hypothesized to have potential to offer additional insights. Therefore, sites from the early (N ¼ 500) and middle Holocene (N ¼ 1160), including all known offshore sites, were tested for correlations to landscape features including fluvial features, geologic formations providing access to high-quality tool stone, coastline positions from 15 KYA to 5 KYA, and net primary productivity estimates (a proxy for crude biomass availability) (Fig. 54.3). First pass regressions failed completely, with various scores indicating heteroscedasticity and spatial non-stationarity – in short, variation in explanatory power for variables across the study’s spatial extent. Use of individual hotspot analysis to detect clustering around specific landscape variables was more successful, but multivariate statistical analysis of clusters revealed a significant problem: sites correlated to different environmental variables both spatially and temporally. Specific site usage, changes in foraging strategy, population increases, increasing regionalization of cultural practices, and possible population reorganizations can be offered as explanations for this variation, with no resolution for this

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1615

equifinality. Even more problematic were site distributions that, from a strictly evolutionary perspective, were not adaptive. During the early Holocene, waterstressed areas retained significant populations, and past measures of biomass remained significant predictors for clustering. By the middle Holocene, none of the correlations between landscape variables and site locations were, statistically speaking, robust. In short, human decision-making across this landscape varied not only spatially and temporally, but along apparently cultural axes without direct linear relationships to extant ecology [132].

54.3.5 Locational Analysis Models in New Zealand Settlement models that consider the relationship between places, polities, and economic variables make the underlying assumptions that settlements are sedentary and that political boundaries are stable over time [48]. The patterns of sites, which are the basis for these techniques, also rely on a lack of chronological change and assume continued settlement in the same locations over the time period being considered. These standard models were the basis for a test case in an area of New Zealand. The study area is approximately 25 km long and 5 km wide, bordering the lower Waihou River, in the northern North Island (Fig. 54.4). This was an untypical location for Maori settlement, as it was not a good area for cultivating kumara (sweet potato, Ipomoea batatas), yet at the time of European contact (AD 1769), it was one of the most powerful regions in the northern half of the North Island, and the population may have been around 2000 people [136], although other authors suggest that there may have been three times that number [137][138]. One of the settlements in particular, Oruarangi, has yielded the greatest number of portable artifacts found in a Maori site, and archaeological and other studies have described a complex history of settlement along the river [136,137,138]. A significant number of Maori settlements were constructed along the low-lying riverbanks, and a series of fortifications (pa) and small undefended villages (kainga) have been archaeologically recorded there [136]. These settlements were probably occupied between AD 1450 and 1850 (all but one of the radiocarbon dates from nine sites date between 1550 and 1750 cal AD at 1 s.d.). Concentrations of settlement occurred at the confluences of the major eastern tributaries with the Waihou River. Among these was the extremely large fortification (2 ha) of Oruarangi, which was between two and eight times greater than the others. Models such as central place theory would interpret the fortifications along the Waihou River as the principal sites with the smaller villages as satellites, ranksize rule would assume that Oruarangi was the principal site, while Thiessen polygons and catchment analysis would divide the riverbanks into economic and political territories based on the centrality of the defensive settlements. These models would suggest that the larger distribution gaps related to spaces between polities [48].

1616

O. Grøn et al.

Fig. 54.4 Maori settlements along the lower Waihou River, Hauraki, dating from AD 1450–1850 [48]

54.3.6 Evaluation of Locational Analysis Models Along the Waihou River, New Zealand Maori oral accounts along the Waihou River, dating to events prior to and just after European contact, describe a dynamic and complex system of land use was practiced [136]. This system is significantly at odds from the static one assumed by locational models.

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1617

These accounts describe changes based on the shifting gardening rounds whereby trees and shrubs were burnt to produce potassium that enabled the cultivation of kumara for a period of 3 years, after which the potassium would be exhausted. The inhabitants would then have to shift their gardens and often their villages (kainga) to a new area and begin the process again. The new locations would be within a generally accepted sub-tribal (hapu) territory and depend on rights inherited from either parent or marriage partner. The territories of sub-tribes overlapped to some extent and also with the rights to collect or hunt natural resources, such as birds, tree fruits, fish, shellfish, timber, and fiber. Some of these use rights were limited to particular individuals or families, while others were available to much wider groups. Such rights were kept alive through re-occupation and re-use by members of the group: a system known as ahi ka (literally, keeping the fires burning). Maori resource rights along the Waihou River are best described as an overlapping patchwork, which the inhabitants negotiated through their genealogical and historical relationships. Fortifications were built by one or more sub-tribes when there was a threat of attack and would be maintained while the threat remained. In times of peace, the defended sites would be left to decay, although on some occasions they would be sustained in order to enhance the status of the chiefs and used as ceremonial centers. This fluid system of occupation meant that the location of defenses, houses, and gardens changed frequently, and the focus of settlement might change over short periods of time, depending on environmental, economic, social, and political factors. This meant that there were no stable territories and no sedentary housing, but there were many small-scale changes over time. Along the Waihou River, multiple small-scale changes over this 400-year time period were mediated by cultural concepts such as ahi ka, whakapapa (descent), mana whenua (political and spiritual control over the land), and take (cause or reason for action). Similar concepts have been mentioned elsewhere in New Zealand as being fundamental to Maori land use [139: 25–29], [140:115–123]. This case study tests the common geographic and ecologically derived spatial analysis models and demonstrates that they cannot be used on such flexible types of settlement. A very different model is shown in Fig. 54.5. This model demonstrates that over time a dense pattern of villages and defenses can develop, but the relationship of places to each other and the people cannot be identified spatially due to the fluidity of the land-use system. In fact, any use of this type of analysis can only be suitable where it is known that occupation was stable over a considerable period, where the housing was sedentary and the boundaries were fixed.

54.3.7 Agent-Based Modelling (ABM) in the Southern Hebrides Mesolithic Project (SHMP) Agent-based modelling was used in the Southern Hebrides Mesolithic Project running from 1988 to 1989 to model Mesolithic foraging and thus settlement

1618

O. Grøn et al.

Fig. 54.5 Model of movement over three phases at 3–10-year intervals, compared to the palimpsest that archaeologists discover in which short-term movements and identities are unknown. Top: settlement movement of two Maori hapu (shaded orange and purple squares and circles). Base: the palimpsest of the same settlements (shown in black) [48]

patterns in the Southern Hebrides [141]. In a study published in 1999, the “MAGICAL” (ABM) software was used to test to which extent the availability of hazelnuts influenced the movement of Mesolithic foragers thought to be landing on the coast of the island of Islay and staying there [141, 142]. The MAGICAL software can use agents equipped with different user-specified action and decision tokens (“geno-types”) that are updated along with their subsequent roles in decision-making. Each agent has its own cognitive map linked to a GRASS GIS package relating it to geographical data, and experienced landscape information is shared between the agents [142]. First, a model of the relative abundance of hazelnuts in Islay in the Mesolithic was developed taking into account the stipulated relative density of hazel trees as well as factors that cause variation in their productivity. Second, a very simple model of

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1619

foraging was developed consisting of four aspects of foraging for hazelnuts: explicit decision-making (most often relying on optimal foraging theory – OFT), the exchange of information, a seasonal round, and changes of base camp. It was assumed that the Mesolithic agents were exploring uncharted territory and over time gained experience with the landscape. It was assumed that the annual round involved a spring and a summer dispersal, still facilitating local knowledge exchange, but group-wide information exchange at the base camp level only appeared in the hazelnut season during which the group voted on whether to change base camp at the end of each month. Three potential landing places for the group are assumed [142]. The result of this exercise is a poor fit between the registered Mesolithic finds and the simulation output. It, thus, seems unlikely that the distribution of Mesolithic flint artifacts on Islay can be explained in terms of small groups landing at the coast and then foraging for hazelnuts [142].

54.3.8 Evaluating the Agent-Based Modelling (ABM) Employed in the Southern Hebrides Mesolithic Project (SHMP) The use of ABM (MAGICAL software) to distinguish foraging patterns and, thus, settlement patterns in the Southern Hebrides Mesolithic Project takes at the methodological level a lot of features into consideration in addition to the topographical data such as vegetation, decision-making, the development of experience, intragroup interaction, and exchange of information [142,143]. One problem with such advanced systems is, however, that the results they produce will tend to be less accurate than the most imprecise data they are fed with. And one has to feed them with a lot of data to make them produce results. The reconstruction of the relative availability of hazelnuts is based on very little evidence and a lot of modelling that in principle can be quite far off the real prehistoric situation because of competition factors between different tree societies or other complicating features that are difficult to reconstruct [142]. As already mentioned (Sect. 54.2.2.2), the different plant societies and animal societies tend to display spatial-temporal variation over time, at the small-scale level, as well as in relation to each other, why their reconstruction in relation to specific prehistoric situations becomes a challenge in itself. The hazelnut foraging model is so simplified that it is unrealistic. Of course it makes everything easier to model if the focus can be restricted to one resource. On an annual basis, that, however, must be regarded as highly unrealistic. Optimal foraging theory assumes that hunter-gatherer groups have as their sole aim to get as much food value out of their movements as possible. That is a serious underestimation of their high level of insight in their environment, expertise in monitoring its spatiotemporal variation, and knowledge of how to utilize and manage it with the least effort [91:41–49], [92, 93, 144, 145:440–441]. The seasonal round and the annual group variation could as well be totally different from what is assumed, why this is another weak link in the chain.

1620

O. Grøn et al.

The more factors we want ABM systems to take into account, the more low-quality data we have to feed into them with significant consequences for the “correctness” of the results we obtain from them. To acquire high-resolution and high-quality data sets, where possible, would be immensely expensive, and still some important data would be impossible to acquire with sufficient spatiotemporal resolution such as a precise reconstruction of the relevant prehistoric environments with their small-scale dynamics. An issue in the discussion has been whether the type of rationality/logic characteristic for our logical thinking and therefore central in the development of ABM can also be applied to the prehistoric cultures’ decision-making and behavior [144]. Here, one should include behavioral and ideological rules such as avoiding to kill and consume the groups’ different totem animals and rules influencing different cultural subgroups’ settlement locations in different ways [29,146,147]. In prehistoric situations where there is no information preserved of such rules, one will have to distinguish them as part of the data analysis and modelling – if possible. The problem of necessary simplification versus realism in ABM-based socialenvironmental archaeological simulations is an inherent methodological problem that is unlikely to be solved satisfactorily ever [148,149]. One must accept that the models created with ABM cannot be regarded as “real” before they have been verified as matching concrete observable archaeological data features by analysis/ testing or at least have been substantiated through critical analysis of their structure and consistency [144, 149, 150].

54.4

Examples of Complex Landscape and/or Cultural Features

54.4.1 Hidden Assumptions It is important to be aware what assumptions are made in modelling. Often an inherent assumption is that the resources are stable and evenly distributed in the landscape, which is very rarely the case (see Sect. 54.2.2.2). The fishing-site model models, without being very explicit about it, potential settlement zones on the coast, so it is not a method for distinguishing submerged Late Mesolithic inland sites on rivers and lakes. Using it as the only method for locating submerged sites thus includes the hidden assumption that only coastal sites are the target even though quite significant inland sites are known from that period [35, 151:198–202]. It also assumes that the coastal sites it targets have a total focus on marine resources (see 54.3.1). However, it is well known that land resources played an important role also at the coastal sites in the Late Mesolithic [152, 153:293–301]. This may have played a role for that their positions often diverge from those prescribed by the model [29]. It would be logical to expect an annual rhythm of varying availability of the different available resources in the coastal zone that could cause the optimal resource positions to vary through the different seasons. The model does not take such a perspective into account.

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1621

This section presents some examples of phenomena that are often not taken into account in relation to prehistoric topographies/bathymetries. The idea is not to present a full catalogue but to give an idea of what types of physical and cultural features that may have made the prehistoric landscape use much more complex than its topography/bathymetry immediately indicates. The features of landscapes can vary over the year, so they offer very different opportunities during the different seasons (e.g., 4.2). Different parts of a landscape can be more or less attractive to the prehistoric cultures during different periods without any obvious reason. Example 3.2 focusses on a quite restricted area in Northern Jutland, Denmark, which was apparently quite interesting/attractive for the Late Paleolithic Bromme Culture 17 of its sites registered. No Mesolithic sites have been registered in the same area. Example 3.3 deals with the changing interest focus in the topographically flat Åmose basin from the Early and Middle to the Late Mesolithic. Watercourses and lakes can play an important role for their local environments, creating “pockets” that deviate from their surroundings. Section 54.3.4 describes the effect of deep lakes as environmental buffers and “naled” areas as “resource magnets.” In 2.4.2, it was discussed whether the formal logic employed in our culture could be used to model the decision-making of prehistoric cultures with a different operative logic, which must be assumed often to have involved ideology [29,146,147]. Section 54.3.6 discusses a rather simple example of this with some implications.

54.4.2 Seasonally Variable Landscape Features: An Example Taute discusses the annual variation of accessible water in a prehistoric permafrost landscape at Deimern, Northern Germany. A number of Late Paleolithic sites from the Hamburg and Ahrensburg Cultures are located in an approximately 500-m-long erosion valley in a hillside. They appear to be related to a little water course fed in the summer by water from the top of the melting permafrost [154:19–20]. Such permafrost phenomena are well known [155]. A consequence is that the settlement conditions in such landscapes will be significantly different during summer and winter. Similar “high” positions for Late Paleolithic sites are known from other areas related to the upper parts of small erosion valleys such as it for instance at the localities Hollendskær for sites of the Bromme Culture and from Jels for sites of the Hamburg Culture (Figs. 54.6 and 54.7), whereas other sites from the same phase are placed much lower in the landscape, directly on the banks of the former water systems [156,157]. The former – “higher” – group of Late Paleolithic sites should logically be expected to represent sites from the warmest part of the year, summer sites [154]. Modelling of site positions in prehistoric permafrost areas will have to take into account such annually varying landscape features. The later inland Mesolithic sites, which did not have to cope with a permafrost situation, tend generally not to be placed in such high positions in the landscape but to be closely related to the banks of the permanent water systems, or on islands in them, in some cases even placed on moist floating islands attached to the banks [158–160].

1622

O. Grøn et al.

Fig. 54.6 The Late Paleolithic sites of the Bromme Culture at Hollendskær, Northern Jutland, Denmark, and their approximate heights above the prehistoric water level/floor level of the shallow lake or valley – 1: Nørre Ramsgård, 7.75 m, 2: Ramsgård V, 8.25 m, 3: Ramsgård IV, 6.25 m, 4: Ramsgård III, 8.25 m, 5: Ramsgård IX, 9.00 m, 6: Ramsgård II, 8.25 m, 7: Ramsgård VII, 9.50 m, 8: Ramsgård I, 8.25 m, 9: Ramsgård VI, 9.50 m, 10: Varbro I, 6.00 m, 11: Varbro IV, 6.00 m, 12: Varbro II, 6.25 m, 13: Varbro III, 6.25 m, 14: Varbro V, 11.00 m, 15: Kærgård Syd, 19.25 m, 16: Houbak, 16.25 m, 17: Kærgård Nord, 14.25 m. Twelve of the 17 sites are related to features that look like old water eroded ravines [161]. The national Danish cultural heritage database: http:// www.kulturarv.dk/fundogfortidsminder/)

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1623

Fig. 54.7 The Late Paleolithic sites of the Hamburg Culture, Jels 1 and Jels 2, both approximately 7 m above the floor of the valley with the Jels Lakes. The remains of eroded ravines can be seen related to both [162:17–22]. The national Danish cultural heritage database: http://www. kulturarv.dk/ fundogfortidsminder/)

It should be noted that there in the Hollendskær area with 17 Late Paleolithic sites registered so far, no Mesolithic sites have been registered. Apparently, there seems to have been a good reason for the Late Paleolithic but not for the Mesolithic hunters to be there. Annual variations in landscapes that affect the settlement patterns are not restricted to permafrost areas. Annual periods of flooding, the presence of snow during winter, which can be melted and drank far from the water systems [45], etc., can cause significant annual variations in the settlement patterns. Different parts of a landscape can also have different attraction for the prehistoric cultures in different chronological phases.

54.4.3 Variation Over Time in an Area with Very Little Topographical Features The Åmose basin in Zealand, Denmark, is a good test case for Mesolithic site distributions. The Åmose basin measures approximately 4  10 km [163:225]. Its surface is flat with a few moraine features showing some meters above the present peat surface. These features have become more pronounced during the commercial scraping off of up to 1 meter of peat in the basin from the 1940s and into the early

1624

O. Grøn et al.

1950s. During this period, the basin was surveyed by the National Museum of Denmark with a careful registration of all the concentrations of white bone and flint in the dark peat [164:10–15]. Because the Åmose basin, thus, has one of the best registrations of sites, it is well suited for studies of Mesolithic settlement patterns. The peat basin has (and had) a flat surface, which, for topographical modelling of site positions today, is rather constant. Knud Andersen shows a series of distribution maps showing the sites from the different Mesolithic phases he distinguished. His “Maglemose” and “handle core” phases are here merged into Maglemose. It is known that the basin in the Mesolithic consisted of a number of smaller connected lakes in process of filling in, but the exact configuration of the basin in its different Mesolithic phases is not known [163:220–225]. The vegetation in the dry land areas surrounding the Åmose basin has followed the normal development through the Danish Mesolithic [164:376–407] but is difficult to relate to any vegetational factors, making certain parts of the peat basin itself better to settle in than others. From a topographical modelling point of view, there is very little basis for assumptions about changes in the site distributions in the basin. However, if one places a vertical straight (red) line as shown in Fig. 54.8, it becomes obvious that there is a considerable change in the way the settlements are placed in the Early and Middle Mesolithic Maglemose phase and in the Late Mesolithic Ertebølle phase. In the former, 66% of the sites are registered to the west of the red line and 34% to the east of it (Σ ¼ 80). In the latter, it is exactly the opposite (Σ ¼ 95) (Fig. 54.8). From a clear main concentration in the basin’s western end, the sites move to a main concentration in its eastern end. Taking into account that this is very little related to any topographical features of the basin or to any known changes in its water system configuration or vegetation, it must be recognized that it will not be possible to model such a change in the site distribution pattern. A precondition would be a high-resolution analysis of the environmental development within the basin, which, in spite of the existing data, would be extremely difficult and expensive to carry out. And even this would not guarantee the acquisition of data that could facilitate a correct modelling. The central problem in this example is that today’s observable topography apparently is not very important for the site distribution over time and that sufficient high-resolution environmental data for explaining/modelling the development will be extremely expensive and time-consuming, will be potentially impossible to obtain, and may not even “explain” the phenomenon if it is not environmental.

54.4.4 Lakes and Rivers: The Effect of Deep Lakes as Local Environmental Modifiers and the Effect of “Naled” Areas as Resource Magnets Lakes and rivers can have a significant impact on their surrounding environment. The Nichatka Lake in the Northern Chita County in Siberia is up to 100 m deep, 28 km long, and up to 2.7 km wide [165]. Due to its water masses, the average

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1625

Fig. 54.8 The sites from the Maglemose phase (upper) and the Ertebølle phase (lower) registered by Knud Andersen in the Åmose Basin. The red vertical line separates the basin into a western and an eastern part with respectively 66% and 34% of the Maglemose sites (Σ ¼ 80) and respectively 34% and 66% of the Ertebølle sites (Σ ¼ 95). (Based on Andersen [163:220–225])

winter temperature in this valley system is 22  C in contrast to 32  C in the surrounding low areas. For valleys without lakes of significant size, this effect is considerably less (Fig. 54.9) [166:16]. This means that the Evenk Ildinow clan living on the Sen River, running to the north from the lake’s northern end utilizing a territory at the lake’s northern end, has an extremely good winter hunt because reindeer, elk, and other game are attracted by the moderate winter temperatures [45] [Grøn’s field notes]. North of Lake Nichatka, around the Sen River, is an extensive “naled” area. “Naled” means ice in Russian. “Naled areas” or “naled ice shields” is the

1626

O. Grøn et al.

Fig. 54.9 Variations in average winter temperature in the northernmost part of Chita County, Siberia, based on Kulakov et al. [166:16]. This area has mountains of up to 2–3000 m in height. In winter, the 100-m-deep Lake Nichatka and its valley system attract so much game, due to its relatively mild climate, that the Evenk hunter-gatherers living here are more or less sedentary

geological term for lakes or rivers that stay frozen to the bottom until late summer in spite of normal summer temperatures around them – for instance, easily +30– 40  C in inland Siberia. A precondition for the preservation that late in the year of the ice in such naled areas is that they are protected from the direct sunlight by mountain shadows. The cold air above the naled area’s ice has the quality that it kills mosquitoes instantly. They are quite small creatures with extremely little insulation. This makes the naled areas magnets for the wild as well as for the Evenk’s domestic reindeer. They can graze on the summerly lake/riverbanks and walk on to the ice when the mosquitoes, which are a significant irritating factor for them, become too bad (Fig. 54.10) [167:field notes]. By placing themselves between Lake Nichatka and the naled area north of it, the Ildinow clan got access to a narrow area with a constantly rich supply of reindeer, summer and winter. The naled area attracts the wild reindeer in the summer as the lake zone does in the winter. In such mountain regions, the local reindeer can reduce

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1627

Fig. 54.10 The naled area “Chinaskaya Naled” in the northern Chita County in the Kalar Mountains with extensive old Evenk settlements on its banks. (Photo mid-summer 1997 by Ole Grøn)

their migrations to up and down the mountains to find the conditions they prefer at all times of the year. Therefore, the Ildinow clan could maintain a permanent base camp on the Sen River, only moving it a few hundred meters up or down the river when the camp area started stinking because of their pollution of it. From an Evenk point of view, it is ideal when you do not have to move camp but can stay, where you are, and have good game coming to you all year round. Such a practice was quite widespread earlier among the Evenk; for instance, the Kaplin clan on the Terteia River in northern Irkutsk County lived in such a permanent “village” of ten tents. A mobile life is only something the Evenk as hunter-gatherers engage in when they have to [167:field notes]. Early contacts note rather permanent settlements of which some are quite large [168:211–215]. A potential explanation of late Paleolithic summer sites at Jels and Hollendskær placed relatively high above the floor of the valley systems they are related to as discussed in Sect. 54.3.2 could be that the valley systems at that time were naled areas that attracted game from a larger area. With summer average temperatures considerably below those of today [164:360], small ice sheets could have been preserved till late summer in such positions. In Hollendskær, the 17 Bromme sites are concentrated around the basin’s southern part, which would have been least sun exposed (Fig. 54.6). If this explanation is correct, it also explains why the area apparently has become much less interesting in the Mesolithic.

1628

O. Grøn et al.

54.4.5 Ideologically and Socially Based Variations in Landscape Behavior Socially based territorial rules are well known to impose restrictions on bands’ use of the landscape in more or less complicated ways [29, 169, 170]. Purely ideological ideas and rules (e.g., taboos) can cause a similar variation in the landscape use of different cultural groups and even between their bands. An interesting case of the latter is the comparison of the location patterns of the winter settlements of the Mistissini Iinuu (formerly the Mistassini Cree) and the Evenk in quite similar landscapes in Canada and Siberia, respectively [29]. The Evenk drink melted snow in the winter – one informant maintained that they preferred the taste of that for water from rivers and lakes. They are, however, very specific that the snow has to be melted properly in a pot or bucket over fire. To eat “raw” even pure newly fallen snow directly during trips is regarded as bad. The Mistissini Iinuu have a taboo on drinking melted snow, why they during winter have to obtain water through holes they maintain in the ice. Consequently, the Mistassini winter camps are always located immediately adjacent to lakes or rivers, while the Evenk winter sites and winter hunting cabins can be found up to 5 km from the nearest water body [29]:Tanner personal communication] (Fig. 54.9). The distance from the Evenk settlements to the banks of the water systems depends on a number of factors. In cold valleys with little wind, where the cold air accumulates on their floors, they tend to place their winter settlements higher than in warmer valleys where downdrafts level out the temperature differences so that the dwellings can be placed somewhat lower. Their main prey animal, the wild reindeer, has similar preferences, why this settlement strategy tends to place the Evenk near their main prey. A crucial factor for the reindeer is, however, that the snow conditions are good for walking as well as that they give access to the underlying vegetation. In cases where this preference overrules their temperature preferences, the Evenk will follow the reindeer and settle where there are good snow conditions for the reindeer [29]. It will obviously be extremely difficult to reconstruct such delicate environmental details, and their potential variation, in prehistory even though they were of major importance to the prehistoric cultures (Fig. 54.11).

54.4.6 Cultural Small-Scale Variation A general assumption in archaeological thinking is that archaeological cultures are typologically uniform, so one can expect the cultural material expression as well as economy to be more or less uniform within the area of a specific culture. Social anthropology, however, provides a much more “difficult” picture [34:173–224] of culturally differing subgroups of small-scale societies where their bands/clans or even individual families for identity purposes distinguish themselves culturally from each other in their material culture, their ornament, language dialects, ritual

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1629

Fig. 54.11 Mistissini Iinuu and Evenk inhabit quite similar landscapes in Canada and Siberia, respectively. Here is a picture of how their settlements would be distributed in one and the same landscape

practices, etc., in a dynamic way [53:55–64],[171],[54:75],[172],[55],[173:1–20], [168:71], [174], [175:12–39],[56][57]. At the same time similarities will typically be signalized as larger scale levels such as the tribal or “cultural”/ethnic group level [176],[177:40–63]. In for instance, agent-based modelling involving the prehistoric societies’ social organization, it is important to include this small-scale level of cultural “autonomy.” As Rapport and Overing [178]:218] put it: Such societies lacked the basic law-and-order organizing institutions of Western society. They had no government to speak of, no law courts, police or armies, and not even the market place as we know it. It was clear that they did not compartmentalize their social life into the distinct and separate institutions that we recognize as kinship, economics, politics and religion. Anthropologists found instead that these peoples used the idiom of kinship to frame most of their activities, including those with political, economic and religious intent (Fig. 54.12).

For Stone Age pastoralists, agricultural societies, as well as some hunter-gatherer societies, more complex social organizations must be expected, but also here smaller subgroups with a rather high degree of independence and thus small-scale cultural variation must be expected to be at play according to the data from social anthropology [171, 179:181–204].

1630

O. Grøn et al.

Fig. 54.12 Tlingit Indians of the American Northwest Coast. Raven guests at a potlatch given by the Wolf 32 clan at Sitka, December 1904. The picture shows the variation in signs marking the individuals’ clan relation, status, and sex within a fraction of the Tlingit group. (Photograph by E.V. Merrill)

What has relieved archaeology from having to deal with such a dynamic cacophony of signals can well be that they mainly are expressed in the organic easy-tomodify (e.g., ornament) but generally not well-preserved part of the material culture as well as in the cultural action (ritual and linguistic cultural variation) of the prehistoric cultures. Knapped lithics is not an obvious material to use for emitting social signals when skin, hide, fur, bark, wood, etc., are available. This can explain its – in relation to the dynamics of “live cultures” – often rather slow and uniform development within larger areas [180]. It is important to be aware that such a band/clan level autonomy also can include a dynamic small-scale variation in the decisions about economic strategies [29, 93, 181]. The Evenk hunters hunting on foot, those who have small groups of domestic reindeer for transportation only, and those who have up to around 70–80 per family (with the majority of these functioning as saved-up capital) live side by side but have significantly different economic strategies. The first can restrict their annual round to a few base camps because they do not have to provide grazing for reindeer. Especially the third group has to keep moving to provide grazing for their reindeer unless they live in optimal areas, such as the Ildinow clan. Evenk hunters interviewed were able to explain immediately how their clans would adapt their

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1631

economic strategies to specific changes they were presented with in the interviews (see 54.4.4) [30]:Grøn field notes]. This concurs with Murray’s central conclusion about the economies in the Palaeoeskimo cultures of the Foxe Basin, Canada [182]: [. . .] it is clear that there was a wide range of economic options for Arctic peoples, despite a perception that this might not have been the case. While it is true that the number of different animal resources in the Arctic may have been limited, the ways in which humans chose to exploit available resources were highly variable. Economies varied temporally, geographically and culturally.

Such a small-scale variation in economic strategy is possibly also what is reflected in the discontinuous appearance of early domestic cattle in Northern Germany [183].

54.5

Discussion

This chapter has discussed and carried out concrete tests/evaluations of a specter of different methods for topographical modelling of potential Stone Age settlement zones. This has been done with an interdisciplinary perspective on the methods and their concrete performance and with a strong focus on their theoretical background. The cultural as well as environmental small-scale dynamics that today are accepted as a normal equilibrium situation in cultural landscapes have been taken as a natural precondition. In the three preceding sections, this displays a lot of methodological problems. The methods used so far appear to have been operating in an artificially created theoretical space where a series of the real world’s significant problems and complexities have been ignored because they are difficult to handle the nmodelling [149]. The simpler types of topographical/ bathymetric modelling have been based on simplistic assumptions about a rather even and stable distribution of the resources in the environment, rather uniform and stable cultural behaviors. The more advanced types of modelling such as agent-based modelling (ABM) demand an input of large amounts of reconstructed landscape data as well as behavioral profiles for the individual agents to operate within the frames of reasonably realistic cultural action and decision patterns. It is difficult to see how attempts to take into account the environmental smallscale variation and dynamics in complex modelling (e.g., ABM) that live huntergatherers utilize actively in their foraging strategy can become reasonably precise, because the data are simply not there anymore, and the best approximations will be extremely expensive and most likely still of a far too bad quality to do the job in a reliable way. The same is the case for dynamic behavioral differences between cultures as well as between cultural subgroups such as bands/clans over time. A central problem is that imprecise data when forming the basis for complex modelling will result in no less imprecise results according to the rules of propagation of uncertainties [184, 185]. In such cases, the uncertainties of the resulting models

1632

O. Grøn et al.

will be impossible even to estimate statistically, even though they necessarily will reach a level that within the technical disciplines would render them invalid even as insignificant pointers. An alternative to calculating the uncertainties of the results of more complex modelling approaches is to test them against independent “real-world” archaeological data as generally suggested in the later literature [186, 144]. As demonstrated in Sects. 54.3.1–54.3.4, however, the results of such tests, in good compliance with the involved rather uncontrollable uncertainties, do not match the real-world data well. The archaeological modellers’ reaction to the environmental and cultural complexities described and dealt with in modern ethnography/social anthropology has been to ignore them [34], [186]. The same appears to be the reaction to modern landscape ecology and geomorphology. That is not a promising start for a methodological approach that has already been widely applied. The only alternative to that attitude is to engage in the application of complex and dynamic landscape and culture data and see how far they can carry. It is a positive sign that testing of the results of modelling against independent archaeological data as far as it is possible is seen as increasingly important in advanced modelling [59, 144, 150]. Even though such testing can be difficult, it demonstrates a will to link up to reality. The central points in such an attempt to incorporate complex environmental as well as cultural modelling will be (1) to find ways of simplifying as much as possible without increasing the accumulation of uncertainty and (2) to develop methods – involving testing – to estimate realistically the potential uncertainty. Whereas landscape ecology is a well-developed discipline with an ongoing development of its ability to model mathematically spatiotemporal environmental dynamics [187, 188, 189] in spite of some structural problems of the same character as those met in archaeological modelling [190], we must accept the fact that our understanding of small-scale societies in social anthropology and especially in archaeology is very far from that level. This can have to do with that humanistic research traditionally does not employ mathematical modelling but can also be due to the fact that Homo sapiens sapiens as two of its central qualifications (1) is able to interact with its natural and cultural environment in an intelligent and dynamic way that is difficult to develop general models for and (2) has a flexible social organization [191] that facilitates a wide range of adaptations allowing for incorporation of highly individual traits, which also are difficult to model on a general basis. In all circumstances, it will be of the utmost importance for progress in this field to develop an ability to better model human small-scale cultures and their adaptive development of strategies for interaction with their physical and cultural environments. Important for such attempts will be their ability to separate differing strategy/ response patterns for the different cultural groups as well as potentially for their subgroups – variation in relation to their social organization or in relation to different landscape situations (as, for instance, observed in Sect. 54.3.4). The access to stronger computing facilities will not in itself provide a solution in itself because it will be practically impossible to feed them with the necessary amounts of data of a sufficient quality.

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1633

In relation to cultures from the colder periods in western Europe, it is a problem that too little is known about their contemporaneous permafrost phenomena such as the small erosion ravines discussed in Sect. 54.3.2 and naled areas as discussed in Sect. 54.3.4. These are far from simple phenomena but apparently quite important to understand in relation to modelling on the basis of the existing topographies/ bathymetries. The alternative to topographical modelling is direct physical detection. This has produced promising results for submerged sites with knapped flint even if they are covered by a couple of meters of seafloor sediment with a preliminary technical setup employing an off-the-shelf chirp sub-bottom system (Teledyne Chirp III – Chap. XX). The development on the basis of this of an optimized system that can also be used for detection of other knapped minerals at submerged sites will, however, take some time. The potential adaptation of this acoustic principle to land situations, where it will be much more demanding to make it work because of the sound insulating effect of air pockets between the sediment grains, will take even more time. It should be noted that even an optimized acoustic system for the detection of knapped lithic material will not be able to distinguish sites with no knapped lithics. Such a technology, like all technologies, has its limits. It is therefore seen as important to attempt to solve the present problems related to topographical/ bathymetric modelling as far as that is possible to open for a future interaction between such approaches and the verification that an optimized acoustic technology can provide. In summary, there is no cheap fix. The examples/tests above show that paleoenvironmental changes have a significant effect on the places where sites/settlements were originally located. Using modern physical data will inevitably result in erroneous conclusions. However, even if past environments are reconstructed in detail, the ethnographic studies outlined above show that social and cultural requirements can at times totally outweigh simple physical requirements. Acknowledgments Parts of this study are funded by the German Academic Scholarship Foundation (Studienstiftung des deutschen Volkes) [AZ] and the White Rose College of the Arts and Humanities (AHRC) [AZ].

References 1. Bettinger RL (1980) Explanatory/predictive models of hunter-gatherer adaptation. Adv Archeol Method Theory 3:189–255 2. Kohler TA, Parker SC (1986) Predictive models for archaeological resource location. Adv Archeol Method Theory 9:397–452 3. Verhagen P (2007) Case studies in archaeological predictive modelling. Leiden University Press, Leiden 4. Brøndsted J (1938) Danmarks Oldtid. Stenalderen. Gyldendalske Boghandel Nordisk Forlag, København 5. Sinitsyn AA (2003) A palaeolithic ‘Pompeii’ at Kostienki, Russia. Antiquity 77(295):9–14 6. Ratzel F (1909) Anthropogeographie. Grundzüge zur Anwendung der Erdkunde auf die Geschichte. Verlag von J. Engelhorn, Stuttgart

1634

O. Grøn et al.

7. Crumley CL (1979) Three locational models: an epistemological assessment for anthropology and archaeology. Adv Archeol Method Theory 2:141–173 8. Freilich M (1967) Ecology and culture: environmental determinism and the ecological approach in anthropology. Anthropol Q 40(1):26–43 9. Jochim MA (1976) Hunter-gatherer subsistence and settlement: a predictive model. Academic, New York 10. Steward JH (1961) The methodology of multilinear evolution. University of Illinois Press, Urbana 11. Williams L, Thomas DH, Bettinger RL (1973) Notions to numbers: Great Basin settlements as polythetic sets. In: Redman CL (ed) Research and theory in current archaeology. Wiley, New York, pp 215–237 12. Wood JJ (1978) Optimal location in settlement space: a model for describing locational strategies. Am Antiq 43(2):258–270 13. Boas F (1896) The limitations of the comparative method of anthropology. Science N.S. 4: 901–908 14. Forde CD (1934) Habitat, economy and society. A geographical introduction to ethnology. Methuen, London 15. Herskovits MJ (1949) Man and his works. Alfred A. Knopf, New York 16. Kroeber AL (1939) Cultural and natural areas of native North America. University of California Press, Berkeley 17. Wallis WD (1926) Geographical environment and culture. Soc Forces 4(4):702–708 18. Wissler C (1926) The relation of nature to man in aboriginal America. Oxford University Press, New York 19. Binford LR (1968) Archaeological perspectives. In: Binford SR, Binford LR (eds) New perspectives in archaeology. Aldine Publishing Company, New York, pp 5–32 20. Shanks M, Tilley C (1992) Re-constructing archaeology. Theory and practice, 2nd edn. Routledge, London 21. Jarman MR, Vita-Finzi C, Higgs ES (1972) Site catchment analysis in archaeology. In: Ucko PJ, Tringham R, Dimbleby GW (eds) Man, settlement and urbanism. Duckworth, London, pp 61–66 22. Kummer H (1970) Spacing mechanisms in social behavior. Soc Sci Inf 9(6):109–122 23. Wilmsen EN (1973) Interaction, spacing behavior, and organization of hunting bands. J Anthropol Res 29:1–31 24. Spaulding AC (1960) The dimensions of archaeology. In: Dole GE, Carneiro RL (eds) Essays in the Science of Culture. In Honor of Leslie A. White. Thomas Y. Crowell Company, New York, pp 437–456 25. Clark DL (1972) Models and paradigms in contemporary archaeology. In: Clarke DL (ed) Models in archaeology. Methuen, London, pp 47–52 26. Hodder I, Orton C (1976) Spatial analysis in archaeology. Cambridge University Press, Cambridge 27. Barth F (1969) Introduction. In: Barth F (ed) Ethnic groups and boundaries. The social organization of culture difference. Waveland Press Inc., Long Grove, pp 9–38 28. Damas D (1984) Introduction. In: Damas D (ed) Handbook of North American Indians. Arctic. Smithsonian Institution, Washington, pp 1–7 29. Grøn O (2018) Some problems with modelling the positions of prehistoric hunter-gatherer settlements on the basis of landscape topography. J Archaeol Sci Rep 20:192–199 30. Grøn O, Turov MG (2007) Resource ‘pooling’ and resource management. An ethnoarchaeological study of the Evenk hunter-gatherers, Katanga County, Siberia. In: Hårdh B, ennbert K, Olausson D (eds) On the road. studies in honour of Lars Larsson, Acta Archaeologica Lundensia in 4o, No. 26. Almqvist & Wiksell International, pp 67–72 31. Sutton P (2010) The logic of Wik camping, North Australia. In: Hardy K (ed) Archaeological invisibility and forgotten knowledge. Conference proceedings, Łódź, Poland, 5th–7th September 2007, BAR International Series 2183. Archaeopress, Oxford, pp 91–107

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1635

32. Woodburn J (1972) Ecology, nomadic movement and the composition of the local group among hunters and gatherers: an East African example and its implications. In: Ucko PJ, Tringham R, Dimbleby GW (eds) Man, settlement and urbanism. Duckworth, London, pp 193–206 33. Lévi-Strauss C (1967) Structural anthropology. Doubleday Anchor Books, New York 34. Renfrew C, Bahn P (2000) Archaeology, 3rd edn. Thames & Hudson, London 35. Fischer A (1995) An entrance to the Mesolithic world below the ocean. Status of ten years’ work on the Danish sea floor. In: Fischer A (ed) Man and sea in the Mesolithic. Coastal settlement above and below present sea level. Proceedings of the international symposium, Kalundborg, Denmark, 1993. Oxbow, Oxford, pp371–384 36. Gaffney V, Van Leusen PM (1995) Postscript-GIS, environmental determinism and archaeology: a parallel text. In: Lock G, Stančič Z (ed) Archaeology and geographical information systems. A European perspective. Taylor & Francis, London, pp 367–382 37. Jasiewicz J, Sobkowiak-Tabaka I (2015) Geo-spatial modelling with unbalanced data: modelling the spatial pattern of human activity during the Stone Age. Open Geosci 7:289–307 38. Stančič Z, Kvamme KL (1999) Settlement patterns modelling through Boolean overlays of social and environmental variables. In: Barceló JA, Briz I, Vila A (eds) New techniques for old times. CAA 98. Computer applications and quantitative methods in archaeology. Proceedings of the 26th conference, Barcelona, March 1998. BAR International Series 757, Oxford, pp 231–238 39. M. van Leusen, H. Kamermans (eds) (2005) Predictive modelling for archaeological heritage management: a research agenda. Rijksdienst voor het Oudheidkundig Bodemonderzoek, Amersfoort 40. Wandsleeben M, Verhart L (1997) Geographical information systems. Methodological progress and theoretical decline? Archaeol Dialogues 4(1):53–70 41. Verhagen P, Whitley TG (2011) Integrating archaeological theory and predictive modeling: a live report from the scene. J Archaeol Method Theory 19(2012):49–100 42. Price TD, Feinman GM (2001) The archaeology of the future. In: Feinman GM, Price TD (eds) Archaeology at the millennium. A sourcebook. Springer US, Boston, pp 475–495 43. Grøn O (2017a) Interdisciplinary reflections on repetitive distribution patterns in Scandinavian Mesolithic dwelling spaces. J Archaeol Sci Rep 18:925–935 44. Groß D, Zander A, Boethius A, Dreibrodt S, Grøn O, Hansson A, Jessen C, Koivisto S, Larsson L, Lübke H, Nilsson B (2018) People, lakes and seashores: studies from the Baltic Sea basin and adjacent areas in the early and Mid-Holocene. Quat Sci Rev 185:27–40 45. Grøn O (2012) Our grandfather sent the elk – some problems for hunter-gatherer predictive modelling. Quartär 59:175–188 46. Grøn O, Boldreel LO, Hermand J-P, Rasmussen H, Dell’Anno A, Cvikel D, Galili E, Madsen B, Nørmark E (2018) Detecting human-knapped flint with marine high-resolution reflection seismics: a preliminary study of new possibilities for subsea mapping of submerged Stone Age sites. Underwater Technol 35(2):35–49 47. Madsen DB (1981) The Emperor’s new clothes. Am Antiq 46(3):637–640 48. Phillips C (2004) GIS and Landscape Analysis, or the cart before the horse? Internet Archaeol 16:23 49. Howey MCL, Burg MB Assessing the state of archaeological GIS research: unbinding analyses of past landscapes. J Archaeol Sci 84:1–9 50. Verhagen P, Nuninger L, Tourneux F-P, Bertoncello F, Jeneson K (2013) Introducing the human factor in predictive modelling: a work in progress. In: Earl G, Sly T, Crysanthi A, Murrieta-Flores P, Papadopoulos C, Romanowska I, Wheatley D (eds) CAA 2012: proceedings of the 40th conference in computer applications and quantitative methods in archaeology (CAA), Southampton, England. Pallas Publications, Utrecht, pp 379–388 51. Santos JI, Pereda M, Zurro D, Álvarez M, Caro J, Galán JM, iGodino IB (2015) Effect of resource spatial correlation and Hunter-Fisher-Gatherer mobility on social cooperation in Tierra del Fuego. Plos One:29. https://doi.org/10.1371/journal.pone.0121888

1636

O. Grøn et al.

52. Swedlund AC, Sattenspiel L, Warren AL, Gumermann GJ (2015) Modeling archaeology: origins of the artificial Anasazi Project and beyond. In: Wurzer G, Kowarik K, Reschreiter H (eds) Agent-based modeling and simulation in archaeology, Advances in geographic information science. Springer, Cham, pp 37–50 53. Barth F (1987) Cosmologies in the Making. A Generative Approach to Cultural Variation in Inner New Guinea. Cambridge University Press, Cambridge 54. Donner K (1915) Bland Samojeder i Sibirien åren l911 – 1913, 1914. Söderström & Co, Helsingfors 55. Grøn O, Klokkernes T, Turo MG (2009) Cultural small-scale variations in a hunter-gatherer society: or ‘everybody wants to be a little bit different!’ An ethnoarchaeological study from Siberia. In: McCartan SB, Schulting R, Warren G, Woodman P (eds) Mesolithic horizons. Papers presented at the seventh international conference on the Mesolithic in Europe, Belfast 2005, vol I. Oxbow Books, Oxford, pp 203–209 56. Wiessner P (1983) Style and social information in Kalahari San projectile points. Am Antiq 48(2):253–276 57. Wiessner P (1984) Reconsidering the behavioural basis for style: a case study among the Kalahari San. J Anthropol Archaeol 3:190–234 58. Bullock S (2014) Levins and the Lure of artificial worlds. The Monist 97(3), Models and Simulations 301–320 59. Lake MW (2015) Explaining the past with ABM: on modelling philosophy. In: Wurzer G, Kowarik K, Reschreiter H (eds) Agent-based modeling and simulation in archaeology, Advances in geographic information science. Springer, Cham, pp 3–35 60. Bernhardson M (2018) Aeolian dunes of central Sweden, Lundqua thesis 84, 117 pp. Lund University, Lund 61. Colbo K, Ross T, Brown C, Weber T (2014) A review of oceanographic applications of water column data from multibeam echosounders, estuarine. Coast Shelf Sci 145:41–56 62. Hansson A (2018) Submerged landscapes in the Hanö Bay. Early Holocene shoreline displacement and human environments in the southern Baltic Basin. Lundqua thesis 85, 146 pp. Lund University, Lund 63. Hansson A, Hammarlund D, Landeschi G, Sjöström A, Nilsson B (2019) A new early Holocene shoreline displacement record for Blekinge, southern Sweden, and implications for underwater archaeology. Boreas 48:57–71 64. Caruso A, Cosentino C, Pierre C, Sulli A (2011) Sea-level changes during the last 41,000 years in the outer shelf of the southern Tyrrhenian Sea: evidence from benthic foraminifera and seismostratigraphic analysis. Quat Int 232:122–131 65. Hansson A, Björck S, Heger K, Holmgren S, Linderson H, Magnell O, Nilsson B, Rundgren M, Sjöström A, Hammarlund D (2018) Shoreline displacement and human resource utilization in the southern Baltic Basin coastal zone during the early Holocene: new insights from a submerged Mesolithic landscape in South-Eastern Sweden. The Holocene 28:721–737 66. Long AJ, Woodroffe SA, Roberts DH, Dawson S (2011) Isolation basins, sea-level changes and the Holocene history of the Greenland Ice Sheet. Quat Sci Rev 30:3748–3768 67. Corner GD, Yevzerov VY, Kolka VV, Möller JJ (1999) Isolation basin stratigraphy and Holocene relative sea-level change at the Norwegian-Russian border north of Nikel, Northwest Russia. Boreas 28:146–166 68. Yu SY (2003) The Littorina transgression in southeastern Sweden and its relation to mid-Holocene climate variability. Lundqua thesis 51, 82 pp. Lund University, Lund 69. Mikaelsson J (1978) Strandvallskomplexet vid Olsäng. Blekinges Natur 1978:37–52 70. Tikkanen M, Oksanen J (2002) Late Weichselian and Holocene shore displacement history of the Baltic Sea in Finland. Fennia 180(9–20) 71. Brader MD, Lloyd JM, Barlow NLM, Norðdahl H, Bentley MJ, Newton AJ (2017) Postglacial relative sea-level changes in Northwest Iceland: evidence from isolation basins, coastal lowlands and raised shorelines. Quat Sci Rev 169:114–130

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1637

72. Wilcke JC (1850) Historiska och physiografiska underrättelser om Landskrona stad och hamn sammandragne a:o 1770. pp Stockholm 73. Reid C (1913) Submerged forests. Cambridge University Press, London 74. Isberg O (1927) Beitrag zur Kenntnis der post-arktischen Landbrücke. Geogr Ann 9: 100–108 75. Berglund BE, Berglund P, Blivik J (1986) En forntida tallskog på havsbottnen i Blekingeskärgården. In: Mattsson L (ed) Blekinges natur. Svenska Naturskyddsföreningen, Karlskrona, pp 40–50 76. Berglund BE, Sandgren P, Barnekow L, Hannon G, Jiang H, Skog G, Yu SH (2005) Early Holocene history of the Baltic Sea, as reflected in coastal sediments in Blekinge, southeastern Sweden. Quat Int 130:111–139 77. Habicht HL, Rosentau A, Jõeleht A, Heinsalu A, Kriiska A, Kohv M, Hang T, Aunap R (2017) GIS-based multiproxy coastline reconstruction of the eastern Gulf of Riga, Baltic Sea, during the Stone Age. Boreas 46:83–99 78. Roper DC (1979) The method and theory of site catchment analysis: a review. Adv Archeol Method Theory 2:119–140 79. Grøn O(2005) A Siberian perspective on the north European Hamburgian Culture: a study in applied hunter-gatherer ethnoarchaeology. Before Farming 2005/1 article 3. pp 30 80. Ingold T (1980) Hunters, pastoralists and ranchers. Reindeer economies and their transformations. Cambridge University Press, Cambridge 81. Plew MG (1996) Introduction. In: Plew MG (ed) Prehistoric Hunter-Gatherer fishing strategies. Department of Anthropology, Boise State University, USA, pp 1–5 82. Suttles W (1990) Environment. In: Suttles W (ed) Handbook of North American Indians. Northwest Coast, vol 7. Smithsonian Institution, Washington, pp 16–29 83. Wiens JA (1995) Landscape mosaics and ecological theory. In: Hansson L, Fahrig L, Merriam G (eds) Mosaic landscapes and ecological processes. Chapman & Hall, London, pp 1–26 84. Hansson L, Fahrig L, Merriam G (1995) Preface. In: Hansson L, Fahrig L, Merriam G (eds) Mosaic landscapes and ecological processes. Chapman & Hall, London, pp xvi–xix 85. Levin SA, Powell TM, Steele JH (eds) (1993) Patch dynamics. Lecture notes in biomathematics, vol 96. Springer, Berlin 86. Innes JB, Blackford JJ, Simmons IG (2004) Testing the integrity of fine spatial resolution palaeoecological records: microcharcoal data from near-duplicate peat profiles from the North York Moors, UK. Palaeogeogr Palaeoclimatol Palaeoecol 214:295–307 87. Archibald S, Staver AC, Levin SA (2012) Evolution of human-driven fire regimes in Africa. PNAS 109(3):847–852 88. Bird DW, Bird RB, Codding BF (2016) Pyrodiversity and the Anthropocene: the role of fire in the broad spectrum revolution. Evol Anthropol 25:105–116 89. Cox MP, Morales DA, Woerner AE, Sozanskil J, Wall JD, Hammer MF (2009) Autosomal resequence data reveal Late Stone Age signals of population expansion in Sub-Saharan African foraging and farming populations. PLoS ONE 4(7):e6366: 1–8 90. Daniau A-L, d’Errico F, Goñi MFS (2010) Testing the hypothesis of fire use for ecosystem management by Neanderthal and Upper Palaeolithic Modern Human Populations. PLoS ONE 5(2):e9157, 1–10 91. Anderson K (2005) Tending the wild. Native American knowledge and the management of California’s Natural Resources. University of California Press, Berkeley 92. Berkes F, Colding J, Folke C (2000) Rediscovery of traditional ecological knowledge as adaptive management. Ecol Appl 10(5):1251–1262 93. Grøn O (2017b) Optimal foraging or not? An Evenk example of resource exploitation. Stratum Plus 2007-1. 319–328, in Russian. English version at: https://www.academia.edu/31660978/ Optimal_foraging_or_not_An_Evenk_example_of_resource_exploitation 94. Anderson DG (1991) Examining prehistoric settlement distribution in Eastern North America. Archaeol East North Am 19:1–22

1638

O. Grøn et al.

95. Fischer A (1997) People and the sea – settlement and fishing along the Mesolithic coasts. In: Pedersen L, Fischer A, Aaby B (eds) The Danish Storebælt since the Ice Age – man, sea and forest. Sund & Bælt, Copenhagen, pp 63–77 96. Fischer A (2011) Stone Age on the Continental Shelf: an eroding resource. In: Benjamin J, Bonsall C, Pickard C, Fischer A (eds) Submerged prehistory. Oxbow, Oxford, pp 298–310 97. Benjamin J (2010) Submerged Prehistoric landscapes and underwater site discovery: reevaluating the ‘Danish model’ for international practice. J Island Coast Archaeol 5:253–270 98. Fitch S, Gaffney V, Thomson K (2007) The Archaeology of the North Sea Palaeolandscapes. In: Gaffney V, Thomson K, Fitch S (eds) Mapping Doggerland. The Mesolithic landscapes of the Southern North Sea. Archaeopress, Oxford, pp 105–126 99. Nyland A. Lokaliseringsanalyse av tidligmesolittiske pionerboplasser. Havvind Paleogeografi og arkeologi. (eds. H. Glørstad & F. Kvalø). Norsk Maritimt Museum – Arkeologisk Rapport 2012:12, 70–96 100. Fischer A (1987) Stenalderbopladser på bunden af Smålandsfarvandet. Miljøministeriet, Skov- og Naturstyrelsen 101. Johansson A (1966) Sydsjællands oldtidsbebyggelse. En foreløbig meddelelse Årbog for Historisk Samfund for Præstø Amt 1964:245–328 102. Johansson A (1999) Ertebøllekulturen i Sydsjælland. Årbog for Nordisk Oldkyndighed og Historie 1997:7–88 103. Fischer A, Sørensen SA (1983) Stenalder på den danske havbund. Antikvariske Studier 6: 104–126 104. Larsson L (1983) Mesolithic Settlement in the Sea Floor in the Strait of Öresund. In: Masters PM, Flemming NC (eds) Quaternary Coastlines and Marine Archaeology. Academic, Wellington, pp 283–301 105. Skaarup J, Grøn O (2004) Møllegabet II. A submerged Mesolithic settlement in southern Denmark. BAR International Series 1328 106. Christensen H, Nielsen AV (eds) (1988) Havbundsundersøgelser. Råstoffer og fredningsinteresser. Storebælt Oversigt. Miljøministeriet, Skov- og Naturstyrelsen, Hørsholm 107. Grøn O (2015) Fiskeplassmodellen, en metode for lokalisering av steinalderboplasser under vann - eller det motsatte? Årbok for Norsk Maritimt Museum 2014:235–244 108. Halligan J, Waters MR, Perrotti A, Owens IJ, Feinberg JM, Bourne MD, Fenerty B, Winsborough B, Carlson D, Fisher DC, Stafford TW, Dunbar JS (2016) Pre-Clovis occupation 14,550 years ago at the Page-Ladson site, Florida, and the peopling of the Americas. Sci Adv 2(5):1–8. https://doi.org/10.1126/sciadv.1600375 109. Waters MR, Keene JL, Forman SL, Prewitt ER, Carlson DL, Wiederhold JE (2018) Pre-Clovis projectile points at the Debra L. Friedkin site, Texas – Implications for the Late Pleistocene peopling of the Americas. Sci Adv 4 110. Waters MR, Forman SL, Jennings TA, Nordt LC, Driese SG, Feinberg JM, Keene JL, Halligan J, Lindquist A, Pierson J, Hallmark CT, Collins MB, Wiederhold JE (2011) The Buttermilk Creek complex and the origins of Clovis at the Debra L. Friedkin site, Texas. Science 331(1599–1603). https://doi.org/10.1126/science.1201855 111. Anderson DG, Faught MK (1998) The distribution of fluted Paleoindian projectile points: update 1998. Archaeol East North Am 26:163–187 112. Stright MJ (1986) Human occupation of the continental shelf during the late Pleistocene/early Holocene: methods for site location. Geoarchaeology 1(4):347–363. https://doi.org/10.1002/ gea.3340010402 113. Stright MJ (1995) Archaic sites on the continental shelf of North America: the effect of relative sea-level changes on archaeological site locations and preservation. In: Bettis EAIII (ed) Archaeological Geology of the Archaic Period in North America. Geological. Geological Society of America, Boulder, pp 131–147 114. Garrison EG, Weaver W, Littman S, Cook Hale JW, Srivastava P (2012) Late quaternary paleoecology and Heinrich events at Gray’s reef National Marine Sanctuary, South Atlantic Bight, Georgia. Southeastern Geology 48(4):165–184

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1639

115. Garrison EG, Cook Hale JW, Cameron CS, Smith E (2016) The archeology, sedimentology and paleontology of Gray’s Reef National Marine Sanctuary and nearby hard bottom reefs along the mid continental shelf of the Georgia Bight. J Archaeol Sci Rep 5:240–262. https:// doi.org/10.1016/j.jasrep.2015.11.009 116. Joy S (2018) The trouble with the curve: reevaluating the Gulf of Mexico Sea-level Curve. Unpublished masters thesis. Department of Anthropology, The Florida State University 117. Environments, Coastal Cultural Resources Evaluation of the Northern Gulf of Mexico Continental Shelf: A report prepared for U.S. Department of the Interior, National Park Service, Division of State Plans and Grants (1977) 118. Gagliano SM, Pearson CE, Weinstein RA, Wiseman DE, Mcclendon CM, Simmons J (1982) Sedimentary studies of Prehistoric Archaeological Sites: criteria for the identification of submerged archaeological sites of the northern Gulf of Mexico continental shelf: A report prepared for the U.S. Department of the Interior, National Park Service, Division of State Plans and Grants 119. Pearson CE, Weinstein RA, Gagliano SM, Kelley DB (2014) Prehistoric site discovery on the outer Continental Shelf, Gulf of Mexico, United States of America. In: Evans AM, Flatman JC, Flemming NC (eds) Prehistoric Archaeology on the Continental Shelf: a global review. Springer International Publishing, New York, pp 53–72 120. Cockrell WA, Murphy L (1978) Pleistocene man in Florida. Archaeol East N Am 6:1–12 121. Murphy L (1990) Natural site-formation processes of a multiple-component underwater site in Florida (Submerged Cultural Resources Unit. National Park Service, United States Department of the Interior, Santa Fe, New Mexico 122. Evans AM (2012) Out of site but not out of mind: Submerged prehistoric landscapes on the Northwestern Gulf of Mexico Outer Continental Shelf. Unpublished doctoral dissertation. Louisiana State University and Agricultural and Mechanical College 123. Stright M, Lear EM, Bennett JF (1999) Spatial Data Analysis of artifacts redeposited by Coastal Erosion: a Case Study of McFaddin Beach, Texas, vol I (OCS Study MMS 99–0068. Herndon, Virginia, U.S.A. 124. Turner ES, Tanner P (1994) The McFaddin Beach site on the Upper Texas Coast. Bull the Texas ArcheolSoc 65:319–336 125. Anuskiewicz RJ, Dunbar JS (1993) Of Prehistoric man at Ray Hole Springs: a drowned Sinkhole located 32 KM Offshore on the Continental Shelf in 12 M seawater. In: Heine JN, Crane NL (eds) Diving for Science....1993: Proceedings of the American Academy of Underwater Sciences Thirteenth Annual Scientific Diving Symposium. American Academy of Underwater Sciences, Pacific Grove, California, 1993, pp 1–22 126. Faught MK, Donoghue JF (1997) Marine inundated archaeological sites and Paleofluvial systems: examples from a karst-controlled continental shelf setting in Apalachee Bay. Northeastern Gulf Mexico, Geoarchaeol 12(5):417–458 127. Faught MK (2004a) The underwater archaeology of Paleolandscapes. Apalachee Bay, Florida, American Antiquity 69(2):275–289 128. Faught MK (2004b) Submerged Paleoindian and archaic sites of the big bend. Florida, J Field Archaeol 29(3/4):273–290 129. Marks BS (2006) Site formation processes and activity areas at Ontolo (8JE1577): a submerged Prehistoric human occupation site in Apalachee Bay. Florida. Unpublished doctoral dissertation, Department of Anthropology, The Florida State University 130. Cook Hale JW, Garrison EG (2017) Spatial statistical analysis of coastal plain Paleoindian site distributions and paleoecology: implications for the search for offshore submerged sites in Georgia. Early Georgia 45(1&2):167–180 131. Cook Hale JW, Hale NL, Garrison EG (2018) What is past is prologue: excavations at the Econfina Channel site, Apalachee Bay, Florida, U.S.A. Southeastern Archaeol 1–22. https:// doi.org/10.1080/0734578X.2018.1428787 132. Cook Hale JW, Ervan G, Garrison G (2019) Climate change, cultural adaptations, and lower coastal plain occupations of Georgia and Florida from the early to middle Holocene:

1640

O. Grøn et al.

Extrapolating spatial trends into the offshore 2019. J Anthropol Archaeol 55. https://doi.org/ 10.1016/j.jaa.2019.101071 133. Miller DS (2016) Modeling Clovis landscape use and recovery bias in the Southeastern United States using the Paleoindian Database of the Americas (PIDBA). Am Antiq 81. https://doi.org/ 10.7183/0002-7316.81.4.697 134. Dunbar JS (2016) Paleoindian Societies Of The Coastal Southeast. University Press of Florida, Gainesville 135. Leigh DS (2008) Late quaternary climates and river channels of the Atlantic Coastal Plain, Southeastern USA. Geomorphology 101(1–2):90–108. https://doi.org/10.1016/j.geomorph. 2008.05.024 136. Phillips C (2000) Waihou journeys: the archaeology of 400 years of Maori settlement. Auckland University Press, Auckland 137. Furey LM (1996) Oruarangi. The Archaeology and Material Culture of a Hauraki Pa. Bulletin of the Auckland Institute and Museum 17 138. Monin P (2001) This is my place: Hauraki contested 1769–1875. Wellington 139. McBurney P (2010) Peter: traditional history overview of the Mahurangi and Gulf Islands Districts. Mahurangi and Gulf Islands districts collective committee in association with Crown Forestry Rental Trust 140. Waitangi T (2008) He Maunga Rongo: report on Central North Island claims, vol 1. Waitangi Tribunal, Wellington 141. Mithen S (2000) The Scottish Mesolithic Prospects and the Rationale for the Southern Hebrides Mesolithic Project. In: Mithen S (ed) Hunter-gatherer Landscape archaeology. The Southern Hebrides Mesolithic Project 1988–98, vol 2. McDonald Institute Monographs, Cambridge, pp 9–37 142. M.W. Lake: MAGICAL Computer simulation of Mesolithic Foraging. In: Dynamics in human and primate societies, T. Kohler, G. Gumerman (Oxford University Press, Oxford 1999) pp. 107–143 143. Lake MW (2000) MAGICAL Computer simulation of Mesolithic Foraging on Islay. In: Mithen S (ed) Hunter-gatherer Landscape archaeology. The Southern Hebrides Mesolithic Project 1988–98, vol 2. McDonald Institute Monographs, Cambridge, pp 465–495 144. Barceló JA, Del Castillo F (2016) Simulating the past for understanding the present. A critical review. In: Barceló JA, Del Castillo F (eds) Simulating prehistoric and ancient worlds. Springer International Publishing, Switzerland, pp 1–140 145. Lee RB (1984) The !Kung San. Men, Women, and Work in a Foraging Society. Cambridge University Press, Cambridge 146. McConnel U (1981) The Wik-Munkan Tribe. Part II. Totemism, Oceania I, 181–206–931) 147. S.A. Tokarev 1966: The problem of Totemism as seen by Soviet Scholars, Curr Anthropol 7(2), (185–188 148. Lake MW (2008) Computer simulation modeling. In: Pearsall DM (ed) Encyclopedia of archaeology. Elsevier/Academic Press, San Diego, pp 1134–1140 149. Lake MW (2014) Trends in archaeological simulation. J Archaeol Method Theory 21:258–287 150. Epstein J (1999) Agent-based computational models and generative social science. Complexity 4(5):41–60 151. Andersen K (1983) Stenalderbebyggelsen i den vestsjællandske Åmose. Fredningsstyrelsen, Copenhagen 152. Andersen SH (1991) Bjørnsholm. A Stratified Køkkenmødding on the Central Limfjord, North Jutland. J Danish Archaeol 10:59–96 153. Andersen SH (2013) Tybrind Vig. Submerged Mesolithic settlements in Denmark. Jutland Archaeological Society, Aarhus 154. Taute W (1968) Die Stielspitzen-Gruppen im nördlichen Mitteleuropa. Ein Beitrag zur Kenntnis der späteren Altsteinzeit. Böhlau Verlag, Köln Graz 155. Woo M-K, Kane DL, Carey SK, Yang D (2008) Progress in permafrost hydrology in the new millennium. Permafr Periglac Process 19:237–254

54

Mapping Stone Age Sites by Topographical Modelling: Problems and. . .

1641

156. Mathiassen T (1946) En Senglacial Boplads ved Bromme. Aarbøger for Nordisk Oldkyndighed og Historie 1946:121–197 157. Salomonsson B (1962) Sveriges äldsta kontakt med Västeuropa. En boplats vid Segebro i Skåne. In: Hamberg PG (ed) Proxima Thule: Sverige och Europa under forntid och medeltid. Norstedt, Stockholm, pp 1–25 158. Andersen SH (1979a) Maglemosekultur. In: Rud M (ed) Arkæologisk Håndbog. Politikens Forlag, Copenhagen, pp 193–197 159. Andersen SH (1979b) Kongemosekultur. In: Rud M (ed) Arkæologisk Håndbog. Politikens Forlag, Copenhagen, pp 166–168 160. Andersen SH (1979c) Ertebøllekultur. In: Rud M (ed) Arkæologisk Håndbog. Politikens Forlag, Copenhagen, pp 72–76 161. Nilsson T (1989) Senglacial bosættelse i Vendsyssel. KUML 1987:47–75 162. Holm J, Rieck F (1992) Istidsjægere ved Jelssøerne. Museumsrådet for Sønderjyllands Amt, Haderslev 163. Andersen K (1983) Stenalder Bebyggelsen i den Vestsjællandske Åmose. Fredningsstyrelsen, København 164. Iversen J (1967) Naturens udvikling siden sidste istid. In: Nørrevang A, Meyer TJ (eds) Danmarks Natur Vol 1, Landskabernes opståen. Politikens Forlag, Copenhagen, pp 345–445 165. Shchetnikov AA (2007) Morphotechtonics of lacustrine basins: the Baikal rift zone as an example. Russ J Pac Geol 1(2):120–129 166. Kulakov VC, Krivenko VA, Sologarev AG, Degtev AV, Perfiljeva EE, Likhanov AS, Sulchenko VV, Snitzarenko NI, Gomboeva NG, BIvalajar GA, Balachova LA, Pudenko JT, Schulizh GE, Silberstein IA, Matveeva LG, Dulenpova BI, Umanskayar NV, Popova OA, Kasich VB, Isakova TT, Ogorodnikova LI, Mironova VE, Korsun OV, Romanova NG, Afonini AV, Kotelnikov AM, Gonchikov CD, Gonchikova VK, Nedechev AA, Dornev CB, Malish OF, Kozlova VN, Chipichubov VI, Yadikin LM, Puzanskii VN, Bagova VZ, Cedova CA, Zadoroshnii VF, Arzamashev PV, Gorlachev VP, Fillipov GG, Starchakova NV, Kovalenok NV, Chipichin AA, Kirillov II, Kovichev EV, Druschnina OM, Moschkina ZV (1997) Atlas Chitinskoi oblast i Asinskovo Buriatskovo avtonomnovo okruga. Federalnayar Clushba Geodesii I Kartografii Rossi, Moskva 167. Grøn O, Holm-Olsen IM, Tømmervik H, Kuznetsov O (1999) Reindeer Hunters and Herders: settlement and environmental impact. In: Gundhus G, Seip E, Ulriksen E (eds) NIKU 1994– 1999. Kulturminneforskningens mangfold. NIKU Norsk Institutt for kulturminneforskning, Oslo, pp 20–26 168. Krasheninnikov SP (1972) Explorations of Kamchatka 1735–1741 (Originally published in English in London 1740). Oregon Historical Society, Portland 169. Leacock E (1969) The Montagnais-Naskapi Band. In: Damas D (ed) Contributions to anthropology: band societies. Bulletin No. 228, Anthropological series no. 84. National Museum of Canada, Ottawa, pp 1–20 170. Rogers ES (1969) Band Organization Among the Indians of eastern Subartic Canada. In: Damas D (ed) Contributions to Anthropology: Band Societies. Proceedings of the conference on Band Organization Ottawa, August 30 to September 2, 1965. National Museums of Canada, Bulletin 228. National Museum of Canada, Ottawa, pp 21–55 171. Bird-David N (2017) Before nation. Scale-Blind Anthropology and Foragers’ Worlds of Relatives. Curr Anthropol 58(2):209–226 172. Grøn O, Kuznetsov O, Turov MG (2003) Cultural Micro-Mosaics – a problem for the archaeological culture concept? In: Bergstøl J (ed) Scandinavian archaeological practice – in theory. Proceedings from the 6th Nordic TAG, Oslo 2001. Institutt for Arkeologi, Kunsthistorie og Konservering, Universitetet i Oslo, pp 342–350 173. Helm J (2000) The People of Denendeh. Ethnohistory of the Indians of Canada’s Northwest Territories. University of Iowa Press, Iowa 174. Rigsby B, Sutton P (1982) Speech communities in aboriginal Australia. Anthropol Forum V (1):8–23. 1980–1982

1642

O. Grøn et al.

175. Shirokogoroff SM (1935) Psychomental complex of the Tungus. Kegan Paul, Trench, Trubner & Co., Ltd., London 176. Enfield NJ (2000) The theory of cultural logic. How individuals combine social intelligence with semiotics to create and maintain cultural meaning. Cult Dyn 12(1):35–64 177. Sapir E (2002) In: Irvine JT (ed) The psychology of culture. De Gruyter Mouton, Berlin 178. Rapport N, Overing J (2000) Social and cultural anthropology. The key concepts. Routledge, London 179. Sneath D (2007) The headless state. Aristocratic Orders, Kinship Society, and Misrepresentations of Nomadic Inner Asia. Columbia University Press, New York 180. Sackett JR (1986) Isochrestism and style: A clarification. J Anthropol Archaeol 5(3):266–277 181. Grøn O (2016) Yakut food producers colonising areas occupied by Evenk Hunter-gatherers: fragments of a process of cultural change caused by migration. In: Melheim L (ed) Comparative perspectives on past colonisation, maritime interaction and cultural integration. Equinox Publishers, Sheffield, pp 73–84 182. Murray MS (1996) Economic change in the Palaeoeskimo Prehistory of the Foxe Basin, NWT. PhD dissertation, University of Alberta, Edmonton 183. Hartz S, Lübke H, Terberger T (2007) From fish and seal to sheep and cattle: new research into the process of Neolithisation in northern Germany. Proc Br Acad 114:567–594 184. Arras KO (1998, September) An introduction to error propagation: derivation, Meaning and Examples of Equation CY ¼ FX CX FXT. Tech.Rep. N EPFL-ASL-TR-98-01 R3 of the Autonomous Systems Lab, Institute of Robotic Systems, Swiss Federal Institute of Technology Lausanne (EPFL). pp 21. https://www.research-collection.ethz.ch/handle/20.500.11850/ 82620 185. Ku HH (1966) Notes on the use of propagation of error formulas. J Res Natl Bur Stand C Eng Instrum 70C(4):263–273 186. Verhagen P (2018) Predictive modelling. In: López S (ed) The encyclopedia of archaeological sciences. Wiley-Blackwell, Malden. https://doi.org/10.1002/9781119188230.saseas0475 187. Clark WR (2010) Principles of landscape ecology. Nat Educ Knowl 3(10):34. pp. 7 188. Howell PE, Muths E, Hossack BR, Sigafus BH, Chandler RB (2018) Increasing connectivity between metapopulation ecology and landscape ecology. Ecology 99(5):1119–1128 189. Synes GNW, Brown C, Watts K, White SM, Gilbert MA, Travis JHJ (2016) Emerging opportunities for landscape ecological modelling. Curr Landsc Ecol Rep 1:146–167 190. Verboom J, Wamelink GWW (2005) Spatial modeling in landscape ecology. In: Wiens JA, Moss MR (eds) Issues and perspectives in landscape ecology. Cambridge University Press, Cambridge, pp 79–89 191. Kamilar JM, Baden AL (2014) What drives flexibility in primate social organization? Behav Ecol Sociobiol 68(10):1677–1692. https://doi.org/10.1007/s00265-014-1776-x

Retrospective Photogrammetry: “Building a Time Machine”

55

Colin Wallace

Contents 55.1 55.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . History, Development, and State of the Art of Retrospective Photogrammetry in Archaeology and Cultural Heritage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.3.1 Archival Data Acquisition and Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.3.2 Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.3.3 Camera Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.4 Applications in Archaeological Documentation and Restoration . . . . . . . . . . . . . . . . 55.5 Error, Accuracy, and Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.1 Measurement Error . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.2 Scales Versus Surveying . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.3 Measurement Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.4 Positional Error and Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.5 Image Quality Error and Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.6 Error and Uncertainty in the SfM Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.7 Image Georeferencing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.8 Image Feature Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5.9 Image Reconstruction Error and Uncertainty . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.7 Moving Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1644 1645 1649 1650 1652 1655 1661 1664 1665 1666 1667 1667 1668 1669 1670 1670 1671 1672 1672 1673

Abstract

Retrospective photogrammetric 3D modeling of archaeological sites and monuments using archival photographs has achieved encouraging results as its possibilities and limitations are addressed. This chapter discusses the limitations posed by both the archiving processes and the use of photogrammetric software with regard to camera calibrations. By eliminating elements that reduce the quality of C. Wallace (*) University of Waterloo, Waterloo, ON, Canada e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_55

1643

1644

C. Wallace

the modeling, maximum detail and accuracy can be achieved. By the time they become available in a database or online, archival photographs have gone through several steps in processing, each of which diminishes the quality of the final photographs and their ability to be successfully used in 3D modeling. Based on retrospective photogrammetric projects undertaken in the spring/summer of 2018, this chapter demonstrates how archival photographs can be brought to a level that allows them to perform in the same manner as a contemporary equivalent. This work uses archival photographs of the American School of Classical Studies in Athens for reconstructing archaeological sites in the Athenian Agora (Omega House) in order to aid in preservation and assess deterioration. With this goal in mind, the accuracy of the 3D modeling is paramount and requires novel approaches for improvement.

55.1

Introduction

Photogrammetry is the science of extracting three-dimensional measurements from two-dimensional images. It involves the use of photographs with overlapping, correlating points that allow the construction of a three-dimensional understanding of the subject. Through the spatial variation (parallax) of different points, the x, y, and z-axis of an object can be documented, measurements obtained, and drawings, 3D modeling, and visualizations can be realized. Currently, photogrammetry is extensively used for the metric documentation of cultural assets since it constitutes a unique tool for acquiring measurable, phototextured 3D models of objects, archaeological excavations, monuments, and landscapes (see indicatively [32, 36, 46, 57, 60]. During the past decade, technological advances in cameras, software, and hardware as well as a significant decrease in costs have made photogrammetric data acquisition and processing faster, more automatic, and relatively low cost [46]. 3D documentation has become ubiquitous in cultural heritage preservation to a point that some researchers have recognized a trend toward the use of photogrammetry by inadequately skilled individuals, which can result in their misuse of the technology, since “metric data without certified quality cannot give the correct information” ([57], 259). Quality and accuracy are very important for all scholars involved in excavation, monument preservation, and protection as well as in other areas of cultural heritage protection and management [14, 29, 32]. This becomes a big challenge when trying to work on retrospective photogrammetry, where the data needed for the model is extracted from historical sources, mainly archives. In fact, retrospective photogrammetry is meant to address the production of metrically accurate 3D models derived from archival photographs, produced with film cameras and predating digital photography combining close range terrestrial photographs, low-level aerial photographs, and any other useful materials. “Older photographs, are more challenging to work with due to archival quality, issues of digitization, technical equipment and film media” ([28], 447).

55

Retrospective Photogrammetry: “Building a Time Machine”

1645

Using non-digital photographs presents its own unique set of challenges but does allow the modeler to reach much deeper into time and create the digital reconstruction of cultural heritage that may have changed or does not exist anymore. Retrospective photogrammetry is more challenging than contemporary photogrammetry, is not as widely practiced, and is scarcely published because with the use of archival data, it is very difficult to reach a level of quality in the results that is of an acceptable standard [28, 47, 57]. Maiwald et al. [28] state that “it is possible to reconstruct geometric information about the depicted objects or the camera position at the time of the recording by employing photogrammetric methods,” yet the purpose (and the real challenge) of retrospective photogrammetry is not only to reconstruct geometric information of an object, monument, or site but also to create a metrically accurate full three-dimensional representation of it at the time it was captured and that this representation can be viewed three dimensionally from all angles and not just from one. While producing contemporary photogrammetry has its own rewards, it is a profound moment when a modeler rotates a model made with archival photos and realizes that what is on their screen is essentially a photograph from many decades earlier that was never taken. An example project referred to throughout this chapter is Omega House in the Athenian Agora, Greece. Omega House is a large fourth-century A.D. late Roman structure added to and developed until the sixth century A.D. on the north-east slope of the Areopagus Hill. It is thought to be the last classical school of philosophy. It consists of 30 rooms with elaborate mosaics and complex water systems. It was excavated between 1969 and 1971 by archaeologist John Camp and has remained open to the elements and natural deterioration since then. Structures include low walls, a bath area, and a nearly intact nymphaeum as well as peristyle courtyards. Due to excellent, comprehensive photographic documentation at the time of excavation including aerial photographs, Omega House is an excellent candidate for retrospective photogrammetric modeling to aid in its conservation and restoration. While the ability to do such work economically and without enormous labor is relatively recent, the underlying development of these techniques has a long history.

55.2

History, Development, and State of the Art of Retrospective Photogrammetry in Archaeology and Cultural Heritage

The genesis of the concept of executing photogrammetry using photographs in an archive was initiated by Albrecht Meydenbauer: the father of photogrammetry [4]. “Meydenbauer first had the idea of applying photographic methods to the recording of buildings in 1858” ([41], 762). Meydenbauer’s photographs were made specifically for the purpose of producing photogrammetric drawings (as it was the standard photogrammetric product of the time). His criteria for a proper archival record included photographs for measurement, related field survey documents, and all other data relating to the building’s history ([41], 762). His focus was on architecture, and he understood that you could save the record of a building by

1646

C. Wallace

photo-documenting it. Meydenbauer could not have foreseen the advent of computers, 3D modeling, or retrospective photogrammetry, but his foresight in archiving metric and stereoscopic data created the pasture in which it could be developed. The idea of using archival photographs not initially intended for the production of photogrammetric drawings and measurements of a monumental structure dates at least as far back as the 1960s when Thompson [44] documented the efforts to restore damaged parts of Howard castle in Yorkshire, England. Disparate photographs taken at different times and with different cameras were utilized in the process. Thompson ([44], 114) notes that “the satisfactory results obtained were in very large measure due to good fortune; and we are left with the very uncomfortable feeling that the possibility of faithful restoration of important monuments after serious damage is largely a matter of luck.” Older aerial photographs have been used in photogrammetry for decades, mainly in geographic contexts for mapping or to measure changes in landscapes due to erosion, natural disasters, and human intrusion [5, 55, 61]. These photos are used more often to produce DEMs (digital elevation models) than producing actual 3D models. The advantage of working with aerial photographs is that they have significant overlap, data on flight plans is often known, and they are facing the subject from the same direction. Cantoro ([9], 117) notes that “Aerial photogrammetry in particular takes advantage of the overlap (usually around 60% forward and 30% sidewise) of large-format imagery to numerically recreate a portion of the earth in a virtual environment.” However, producing a 3D model using only higher altitude aerial photographs does lack oblique views and so is not a true 3D model and is not of a sufficient scale to document archaeological features adequately. In their archaeological modeling using only aerial photographs, Peppa et al. do note that “[d] istortion was observed mostly when only vertical imagery was included” [35]. Thus, oblique view can also serve the purpose of reducing overall distortion. Chandler and Cooper [10] first published on monitoring the development of landslides using archival photography and analytical photogrammetry, and in 1992, Chandler and Clark further developed the approach to incorporate archival photographs in order to document buildings. In both cases, analytical photogrammetry was used, which involves using mathematical and geometric analysis to extract metric data from photographs. A few years later, M. Hemmleb [22] revisited A. Meydenbauer’s archive and successfully experimented on digital rectification of single metric images. One year later, Wiedemann et al. [50] conducted similar work using Meydenbauer photographs of the Stadtschloss and Bauakademie buildings in Berlin in order to deliver linear measurements to architects working on the buildings’ reconstruction [50] A similar study was attempted many years later [33] in an effort to create modeling of the now destroyed Great Umayyad Mosque in Aleppo, Syria. The images used were from both metric and non-metric cameras with known calibrations (a metric camera is specifically designed for photogrammetry with a minimum of distortion and with photographic characteristics that do not change from one photograph to the next [26]). Despite this, and with what seems to be a good methodology, the results are not as good as one would expect, and there is significant distortion in the modeling.

55

Retrospective Photogrammetry: “Building a Time Machine”

1647

The first successful attempt at 3D reconstruction of monumental structure is the colossal Buddha statues in Bamiyan, Afghanistan, which were destroyed by the Taliban in 2001 [20]. The images used included four photos acquired from the Internet as well as three images taken in 1970, also using a metric camera. Using the metric images with manual measurements resulted in the most successful 3D modeling [20]. This study was pivotal in the use of older images combined with automated image matching in order to create a 3D model of an object that no longer exists. Ten years later, in 2014, one of the first attempts to use exclusively non-metric, non-digital photographs in the modeling of a now non-existent site was published: Falkingham’s modeling of the Paluxy River dinosaur chase sequence [16]. The results are visually very good but thwarted by the fact that the photographs were consistently taken from one end of the run. This creates occlusions and prevents true 3D modeling. Falkingham mentions nothing about the accuracy of the model but has made the photographs available on his website encouraging other modelers to attempt it and improve upon his result. In the past decade, experiments with contemporary photogrammetric 3D modeling have become more accepted and adopted in the archaeological community due to the advent of structure from motion (SfM). Full 3D models and point clouds dominate the photogrammetric field since SfM photogrammetry (term evolved from computer vision) appeared [30]. SfM takes photogrammetry to a new level, not only providing points in space to theoretically reconstruct objects but actually allowing them to be visually reconstructed using software, which can use the parallax of recognizable points to visualize three dimensions. The term “structure from motion” might not seem to make sense until one thinks of how we, as humans, view our world. Just as the origins of stereoscopic photography view images from two slightly different perspectives, so too do we view our world. However, as we move around an object, our brain processes those stereoscopic views to produce a three-dimensional image of what we are seeing. At any given time, we are only looking at one stereoscopic view of the object, and it is only in our minds that it exists as a fully three-dimensional object. Thus, the motion of our movement around the object creates structure. “Structure from Motion (SfM) has its roots in the well-established spatial measurement method of photogrammetry, but is becoming increasingly recognized as a means to capture dense 3D data to represent real-world objects, both natural and man-made” [56]. The current state of the art in photogrammetry and SfM is one of modeling that achieves photographic realism and accuracies at a sub-millimeter level. That level of accuracy and realism is being achieved with budgets and labor time that are a fraction of what was necessary just 10 years ago. For example, in 2007, the total budget for the 3D documentation (through laser scanning and photogrammetry) of the Acropolis of Athens and the implementation of a GIS came to 885,390.00 €, while the time required for carrying out the work was 18 months ([31], 12). The labor required at that time to do photogrammetry was significant, as was the cost of the equipment. In more recent years, advancements in software, digital cameras, and

1648

C. Wallace

computing power have produced accurate and visually realistic models of large-scale archaeological sites using little more than a DSLR camera, a drone, and commercially available software, producing the models in a matter of hours or days. These gains in efficiency do not carry over as well into retrospective photogrammetry. The archival materials remain as they were decades ago, and it is only through new methodologies that they can be exploited with improved results. Wilson et al. [51] state that “The software is now robust enough to be able to reconstruct 3D models from low-resolution images from uncalibrated cameras.” While this is true in the simplest reconstructions using archival photographs, the examples that they present are of small sets of photographs producing limited views in 3D. Wilson’s team “decided to investigate the idea of producing partial reconstructions of sites using archived data” [51]. Their examples used between four and 15 photographs and were of subjects with little or no occlusions and in some cases (a relief carving and a wall painting) involved a unidirectional subject and view. While this, as with my earlier work on inscriptions and mosaics [46], shows that it is possible to do, what is needed is for techniques to be employed that produce fully three-dimensional models. In looking at other examples of the use of archival photographs in digital reconstructions, there is a shortage of successful attempts. As Wilson et al. point out, “Examples of the use of archive data in the creation of 3D models of heritage sites are scarce” [51]. Even Wilson’s examples show occlusions, distortions, and a lack of full three dimensionality, which is justifiably attributed to the number and quality of the photographs used but, upon further scrutiny, can also be attributed to the methods used, quality of scans, and a lack of camera calibration. Maiwald et al. [28] produced successful 3D modeling of the front facade of the Kronentor (“crown gate”) of the Dresden Zwinger using archival photographs. In that they are only modeling one view of the building, it would be advantageous to have used architectural elevation drawings to rectify the photographs. The building is still in existence in an unaltered state, and therefore, the project is an exercise in testing the ability of archival photographs to be used in modeling rather than an attempt at reconstruction. However, significant numbers of more oblique photographs as well as photographs of the gate from within the courtyard exist, and it is curious that they chose to only model the front facade. Maiwald et al. [28] state that “the very short baselines between the camera positions lead to small intersection angles and thus to low accuracy of the computed object points” and then go on to point out that “In order to achieve better results, a better imaging configuration is crucial. For this purpose, the images taken from other viewing directions, which could not be oriented initially, must be included as well.” In this case, the use of a site plan, aerial photograph, or survey would have helped. They do use contemporary photographs to assist in the geometry but remove them; reference points into the archival model, adjusting those points that are considered reliable and then removing the contemporary photos. Miranda and Melón [59] were able to achieve great accuracy (4.5 cm error) in the modeling of a wall facade in the graveyard of Labraza, Spain, by using stereoscopic photographs taken in 1996 for the express purpose of producing photogrammetric models [59].

55

Retrospective Photogrammetry: “Building a Time Machine”

1649

Zawieska et al. [53] modeled a fortified settlement in Biskupin, northwestern Poland, using multiple modeling techniques. They used a combination of aerial photographs taken from aircraft, blimp, and balloon with altitudes ranging from 50 m to 3 km. Their models were used to make orthoimagery, and no mention of accuracy is given [53]. They rightly note Remondino and Menna’s observations that “Although this problem has been studied for more than 30 years, many problems still exist, such as the impossibility of completeness automation, occlusions, poor or untextured areas and repetitive structures” [37, 53]. These limitations still plague current photogrammetric practices. Further, with regard to the most recent practices, [57] model a building facade using the full range of materials available including floor plans, elevations, archival photographs, and video frames (digitized from motion picture frames). They claim to have accuracies in measurement averaging 13.4 cm by measuring in comparison to architectural plans. Unfortunately, they do not actually present their model. They conclude that a “critical problem is the lack of camera calibration information and the measurements of specific control points and consequently the control of metric accuracy” [57]. Technology has moved forward, and further levels can be attempted. What stands apart in some of the earlier work is that it consistently involved subjects that were more relief: facades of buildings, the Buddhas, and dinosaur tracks. What retrospective photogrammetry attempts to do is to achieve fully surrounding three-dimensional representations of archaeological elements that can be viewed and measured from any angle. The common element or difference here is that whether the modeling is producing a 2.5 dimensional representation of a facade or a horizontal plain, it is not representing full three-dimensional modeling of that element. This is not to criticize the work but rather to explain that the subjects chosen were of a flatter nature giving themselves to more traditional photogrammetric techniques. Another factor that can cause modeling to become 2.5 dimensional or fail completely is that often an aspect of a site or a favored angle dominates the photodocumentation. For example, Peirene Fountain in Ancient Corinth, Greece, has more intriguing features at its south end, and archaeologists over the decades have continually photographed that aspect of the structure, ignoring other angles. For several years the author has used a variety of archival materials, surveys, site plans, elevations, etc., in order to increase the accuracy of the 3D models produced using archival photographs. While doing so, he has also developed techniques and strategies that are able to produce truly three-dimensional, metrically quantifiable representations of archaeological sites and objects. A significant advance is in increasing the quality and usability of the photographs used [46, 47].

55.3

Methodology

The following is the general flow involved in the production of retrospective models and will be discussed step by step. The process referred to is in reference to the use of Agisoft PhotoScan as a 3D modeling software as it has been found to be the most conducive to retrospective photogrammetry.

1650

C. Wallace

55.3.1 Archival Data Acquisition and Issues The process of data acquisition involves a number of necessary steps right up to the point where model processing begins: discovery, gaining access, acquisition (through digital access or scanning), assessment, selection, and identification. Data acquisition involves several hurdles, the first of which is knowing that the data exists. While many collections have been digitized and catalogued, there are basements full of negatives and drawings in places without the time or resources to preserve them through digitization. For example, in Greece, while the American School of Classical Studies at Athens has a large, easily accessible online collection, other institution’s collections can be more difficult to access. The Canadian Institute in Greece has a downloadable access database listing which photographs they have, but there are no thumbnail images and the descriptions, for the most part, are unusable due to their lack of detail, for example, “101R-16 – Vessel – Lamp?” or “101R-08 – Site – Church?”. The British School at Athens’ website states that their collection is currently being digitized and that requests for access can be made formally but there is nothing to indicate what is in the collection. The Australian Archaeological Institute at Athens has no archive mentioned online. The German Archaeological Institute lists a collection of 140,000 scans, which, while extensive, can be arduous to search. The importance of these varied online representations is that while the collection of one institute (probably the one that did an excavation) might have the largest number of photos for a particular site, another may have a smaller number of photos, but those might be crucial in tying together the larger set into a model. Digitized online resources can present a problem in that a given photo subject, over time, can be referred to with several names. For example, Omega House in the Athenian Agora has been called, throughout its excavation life, House C, Omega House, School C, and Ω House. Since useful materials may be temporally separated, not identifying them by their varied labels can result in missed opportunities. Dozens of aerial photographs can be found by including the word aerial, while other groups may use the word “balloon” or refer to the photographer’s name. There is also the problem of Greek locations having several different names. Corinth can also be Ancient Corinth, Corinthos, Korinth, Korinthos, Archea Korinthos, or Αρχαία Κóρινθoς. When suitable photographs and negatives have been located, permissions for access, scanning, publishing, and public presentation must be arranged. In some cases, permissions may be required from more than one institution. For example, those from the Athenian Agora require permission from the American School of Classical Studies at Athens and from the Hellenic Ministry of Culture and Sports. If access to original negatives is granted, permission may also be required to bring a scanner on site. Data acquisition in retrospective photogrammetry differs from contemporary photogrammetry in that the photogrammetrist has little control over the base data. The decision-making process of the original photographer/archaeologist years

55

Retrospective Photogrammetry: “Building a Time Machine”

1651

before determined what was photographed. While photogrammetry requires that there be overlap between photographs, the original intent of the photographer likely did not take that into consideration. Once a data set has been selected, it is to be subjected to further selection criteria. Maiwald et al. [28] in their selection process remove “drawings, postcards and sketches. . . [and] scaffold[ing].” My own research has found that using drawing, plans, etc., can enhance the ability of the software in creating metrically accurate and visually cohesive geometries [47]. Due to the nature of photographic cataloguing and identification, exact locations where a photograph has been taken or the relationships of multiple photographs can be lacking. This problem falls upon the photogrammetrist who must, if possible, work with site directors, archivists, and photographers to ascertain which photographs belong together. The photogrammetrist, in working with the known photographs, develops a familiarity with the elements and, with more experience with a site, will recognize common elements in other photographs and subsequently be able to add them. In all cases, if scanning access of archival materials is at all possible, it should be pursued. In the process of producing 3D digital models using archival photos, there can be a string of systemic errors introduced, producing less than favorable results, particularly when they are being produced using the photographs from digital archives. First, the photographic image, whether on film or glass, has acquired the distortion of the camera and lens with which it was produced. It also contains the grain of the film or emulsion. That image is then printed onto photo paper using an enlarger. In a photo collection, several enlargers may have been used. Each has its own unique distortions to its lens, and that lens is often not of as high a quality as that of the original camera. The resulting print also has grain, furthering the obscuration of details. The printed photo then sits in an archive waiting for the invention of the digital scanner, which has its own lens and its own distortions. In the case of the archives of the American School of Classical Studies at Athens (ASCSA), a European Union grant allowed for the digitization of large quantities of photographs, but the company contracted to do the work used several different scanners for the sake of time and efficiency. This introduced not only the lens error of the scanner but a diversity of lens errors, compounding the distortion that had accumulated along the way. Scanning of negatives is sometimes done under less than ideal conditions. Extreme heat or humidity can affect the resulting scans. Consistently using a scanner with the same optics is also an important factor in maintaining photo quality, and if there are calibration negatives, they should be scanned at the same location as the original negatives (see section “Camera Calibration”). It is also important in the search for data to inquire if there is archival survey data available. If there is no survey data and the site or stable elements of it still exist, acquiring contemporary survey data can be of immeasurable value. This too will require permits.

1652

C. Wallace

55.3.2 Processing Selection The fundamental process of producing retrospective photogrammetry is very similar to that of regular photogrammetry. The following documents some of the nuances required when using archival photographs. Many software programs have the ability to assess photo quality, and those who have attempted retrospective photogrammetry tend to use this ability as they would in contemporary photogrammetry. However, this does not take into account the very limited number of photographs that one has available for retrospective model, and if at all possible, even borderline photographs should be considered. One poor-quality photograph may be what is necessary to pull together the geometries of several good photographs. This is one of the most important factors to remember when using archival photographs. Techniques that may work with digital photographs may have a negative effect when working with limited numbers of archival photographs. Some of the techniques required may even seem counterintuitive but through experimentation have been shown to work. Once the photographs are selected, it must be taken into consideration that in modeling archaeological sites, photography takes place throughout the excavation process. Photographs may be taken of partial excavation or of excess excavation before backfilling. These photographs are still important and usable. With masking of unwanted characteristics, vital metric data can be included. Workers, buckets, shovels, etc., also need to be masked leaving only the features that will support the modeling. Maiwald et al. [28] did not use photographs that included scaffolding. They also excluded images taken at a distance out of concerns that they would affect image matching results, but once again, the changes in parallax that they provide can be useful and they can then be excluded from the dense cloud processing or even texturing ([28], 448). Once chosen, images and other elements are then incorporated within the software to produce a sparse point cloud. Point Cloud The sparse point cloud is comprised of points within the photographs that are consistent in their lighting and location. Data such as camera calibrations, survey data, and arbitrary reference points is then added incrementally in order to monitor improvements in the results. Then processes within the software are configured to further reduce errors: removal of reconstruction uncertainty, removal of projection uncertainty, and optimizing of non-calibrated cameras. Initial sparse cloud processing brings to light photographs that are not advantageous to the model. In some cases, these will become better when control points are added. It is best to leave them in as they can be removed later. When working with archival photographs, after the initial sparse cloud is produced, reconstruction uncertainty can be assessed, and photographs that are causing problems can be disabled but not necessarily removed. It is often the case with archival photographs that two of a similar perspective will create a different plane of points within the

55

Retrospective Photogrammetry: “Building a Time Machine”

1653

model. Disabling one can allow the other to “cooperate” with the rest of the photographs. The model then is made to produce a dense point cloud. Once the dense point cloud is produced, a mesh is made. The mesh is a vector triangulation of all of the points in the dense cloud. The final step is (depending on which software one is using) to colorize the vertices or, if that is automated, to apply texture, draping the photographs over the mesh. The model can now be made into an orthophoto, exported as a 3D model or imported into a VR environment. Older archival photographs are overwhelmingly not photographed for or suited to photogrammetry. It is only through a series of methodological steps that they can be made to produce an acceptable result. Some authors [16, 51] make little mention of what methods they have used, while others [28, 57] use some but not all of the methods necessary to produce the most accurate models. For the execution of retrospective photogrammetry, all recorded elements need to be considered. If a floor plan or site plan is available, it can be treated as a photograph with reference points placed and then added to the photographs. Elevation drawings can also be used but only if the facade has little or no depth because since elevation is orthographic, the points that are closer or further away than the main facade will present as a distortion. As late as the 1970s, accurately scaled watercolor illustrations were painted of some archaeological features, usually mosaics. These, when combined with photographs in which the mosaics are included, can further rectify the modeling. It is important though that all of these elements be of a very two-dimensional nature. Floor plans, for example, should only have reference points added on the same level of the z axis.

Georeferencing If a site is still extant, a site survey is a major step in increasing the accuracy of the 3D modeling. Even in a site that is partially destroyed, if there are points that have remained consistent between the time of original photography and now, they can be used in a survey to pull the model together. Projects in the Athenian Agora and Ancient Corinth have shown that the addition of survey points for elements that have remained unchanged significantly increases the modeling accuracy [47]. If a site is completely destroyed or backfilled, surrounding features (rocks, buildings, etc.) that have remained consistent can be used to unify the photographs. Including these while masking the actual missing features strengthens the geometry. In cases where surveying is not possible either because of site destruction, inaccessibility, or lack of resources, site plans and floor plans are invaluable in bringing together the accuracy of the modeling. It should be noted that these plans cannot be verified metrically and that, in some cases where there is more than one site plan, discrepancies between plans can be seen. PhotoScan estimates internal and external camera orientation parameters during photo alignment. This estimation is performed using image data alone, and there may be some errors in the final estimates. The accuracy of the final estimates depends on many factors, like

1654

C. Wallace

overlap between the neighboring photos, as well as on the shape of the object surface. These errors can lead to non-linear deformations of the final model. During georeferencing the model is linearly transformed using 7 parameter similarity transformation (3 parameters for translation, 3 for rotation and 1 for scaling). Such transformation can compensate only a linear model misalignment. The non-linear component can not be removed with this approach. This is usually the main reason for georeferencing errors. . . Possible non-linear deformations of the model can be removed by optimizing the estimated point cloud and camera parameters based on the known reference coordinates. During this optimization PhotoScan adjusts estimated point coordinates and camera parameters minimizing the sum of reprojection error and reference coordinate misalignment error. [2]

As found in the surveying of Omega House in the Athenian Agora, survey data was not enough to bring together disparate elements and occluded locations, and manual reference points had to be introduced in the modeling. The most successful modeling is achieved when those manual points are widespread and of varying levels on the z-axis. Condorelli and Rinaudo [57] describe a model using extracted images from archival motion picture film for 3D modeling. There is one important factor in whether or not the film is usable in modeling. If the camera is panning while stationary, then all views are from the same perspective and not helpful in photogrammetry. The vast majority of older documentary films use panning and so are not viable. If the film is made with a moving camera, on a rail or otherwise stabilized for movement, then variable angles of the subject are recorded and are able to be incorporated into the modeling. Producing a 3D digital model by simply using a set of photographs can, with an optimal group of photographs, produce what looks like an optimal result. At this point, the only measure of error is in pixels and the subjective eye of the viewer. With the addition of measured, manually placed control points and/or survey data, that error reduces significantly. The more diverse those points are in x, y, and z-axis, the greater the reduction in error. There is, however, a point in the introduction of extra control points at which improvements not only diminish but can start to reverse. There are several reasons for this. With each added control point, there is a greater chance that, due to the varying quality levels of the photographs, a point can be misplaced. When the software suggests points, if there is any doubt about the accuracy of the placement, it is better to ignore that point. Another reason is that as more points are added, the software will include more photographs in which those points have been added. Zawieska et al. [53] note that “Accuracy improves significantly with the number of images where a point appears. Measuring the point in more than four images, however, leads to less significant improvement.” Photographs that were initially deemed to be of the highest quality and compatibility are added first, and then those of lesser quality are added with the addition of more control points. While this may seem to be a less than desirable result, these additional photographs can sometimes help to unite higher-quality elements, resulting in a more comprehensive model.

55

Retrospective Photogrammetry: “Building a Time Machine”

1655

Optimization and Gradual Selection According to Alexey Pasumansky of Agisoft Technical Support, “Error in pixels shows average reprojection error for tie points present on each photo. Usually it should be under 0.6–0.8 pixels. If the error value is over 1 pixel, there are likely some problems” (personal communication). Within Agisoft PhotoScan, there is the ability, once a sparse cloud is constructed, to reduce the reconstruction uncertainty (bundle adjustment) in gradual selection. This removes points in the point cloud with higher uncertainty levels. Part of this process is choosing a numeric level of reduction (with a factor of 10 being recommended), while there is also the ability on the user’s part to see that areas which were already in doubt have been selected for deletion. Once reproduction uncertainty has been reduced, a second reduction optimizing projection accuracy within gradual selection further enhances the accuracy of the results by removing more dubious points. It is at this point that the user can choose to optimize the cameras. Within Agisoft’s programming is the ability to estimate camera error and correct for it. In this way, an approximation of camera calibration is achieved even when lens error is unknown. In order to produce the most reliable and accurate models, we need to remove as many systemic errors as possible. For example, in the summer of 2018, the author visited the archives of the Gennadius Library, Athens, and of the ASCSA in the Athenian Agora and Ancient Corinth and, with the kind permissions of the ASCSA and the Hellenic Ministry of Culture and Sports, directly scanned the original negatives of many archival photos using an Epson V850 Pro. This removed several systemic errors that would normally occur in this process. It also provided the opportunity to adjust the exposures of the photographs to more closely match one another. Many of the scans currently in the digital archive have vastly different exposures (Fig. 55.1a, b). If there is a significant difference in image exposure, successful photogrammetry is less likely. While these can be corrected in Photoshop, Corel Photopaint, or GIMP (Fig. 55.2a, b), detail is lost and so scanning similar exposures at the time of scanning is superior. This process does still leave the problem of the distortion of the original camera.

55.3.3 Camera Calibration All camera lenses have a certain amount of distortion which increases toward the edges of the lens. Some current photogrammetry software includes catalogues of calibration corrections for known contemporary digital cameras. Metadata embedded in the photographs taken with these cameras identifies camera, lens, focal length, and settings at the time of exposure. Corrections are then applied to the photographs during modeling, reducing overall error. Some software will not allow a model to be produced if that information does not exist. Agisoft PhotoScan differs in that it analyzes the photographs and if no calibration information is given, it will estimate the distortion and correct for it. “APP

1656

C. Wallace

Fig. 55.1 (a and b) Differing exposures of scans in the ASCSA photo archive

Fig. 55.2 (a and b) Contrast and lightness corrected archival images

[Agisoft Photoscan Professional] may compute the unknown camera calibration parameters using a self-calibration approach based on tie points extracted during SfM” [6]. Film cameras do not have embedded information, nor do they have calibration data readily available. In such cases, PhotoScan is able to estimate adjustments. Condorelli and Rinaudo ([57], 262) suggest that by adding just the proper focal length, uncalibrated photographs can increase the software’s ability to estimate and thus increase accuracy. By adding this step, there is an increase in accuracy, and it is of little significance. There is, however, the capability of manually inputting calibration data. The following is an account of the author applying calibration to archival photographs and the resulting increase in modeling accuracy. Low-level aerial photographs of Greek archaeological locations became popular in the 1960s and 1970s. For several seasons (stability, control, and very low altitude to capture detail), J. Wilson Myers and Eleanor Emlen Myers used a blimp to photograph archaeological sites in the Mediterranean, primarily using a Hasselblad

55

Retrospective Photogrammetry: “Building a Time Machine”

1657

EL 500 with a Carl Zeiss Distagon Lens 1:4 F ¼ 50 Mm ([48], 352). In 1976, they moved to “tandem camera systems radio controlled from the ground” ([49], 181), but the specific sites being looked at by me were photographed when only the Hasselblad was being used. Inspired by the Myers, James Wiseman of Boston University had a duplicate of the blimp built and used it to photograph the Acrocorinth in Ancient Corinth (Greece), Nicopolis (Greece), and Stobi in Greek Macedonia, also using an identical Hasselblad EL 500 with an identical Carl Zeiss Lens. The Phokis-Doris Expedition [24], also sponsored by the Whittlesey Foundation, involved the aerial “mapping” of “an Ancient trade and communication route which extends 60 kilometers through the mountains of central Greece” ([24], 1489). This method allowed the team to map a route in terrain so rugged that the logistics and expense of conventional surveying would have proved prohibitive. Once again, they were using a Hasselblad EL 500. The relevance of this similarity will become apparent. Initial work using online scans of the Myers’ photographs of the Athenian Agora as well as Ancient Corinth proved successful, and applying retrospective photogrammetric techniques, details from close range terrestrial photos were able to be combined in the models [47]. Low-level aerial photographs taken with film cameras have a couple of limitations when applied to photogrammetry. Firstly, the limitations of shooting with film mean that there are much fewer photographs to work with. Secondly, those photographs were taken with the intent of capturing the most orthographic views of the archaeological sites. The result is that the most oblique views of vertical features in those photographs inhabit the peripheral edges of the photographs where the greatest distortion is manifested. In archives such as that of the American School of Classical Studies at Athens, each photograph represented online has a unique identifying number. These numbers correlate to numbered sleeves containing the negatives in the physical archive. What was unexpected was that upon accessing the actual negatives in the archives of the Athenian Agora and of those in Ancient Corinth, each sleeve contained several negatives under the same number, and the best one was chosen for that particular shot. The others may have not been included because of focus or exposure problems or, in some cases, wind blowing the dirigible and causing the camera to take a more oblique photograph. While these “rejects” were unacceptable for their original purposes due to their lack of orthographic positioning, they capture more details of the sides of elevations and make them better candidates for photogrammetric inclusion. In the archives of the archaeological sites studied, three types of cameras were used: large format, medium format (Hasselblad EL 500), and 35 mm. Large format cameras cannot be calibrated because their bellows can be twisted, and so the variations in focal plane are infinite. It did seem reasonable, however, to assume that calibration could be applied to the medium format and 35-mm photographs if identical cameras could be located and calibration negatives taken.

1658

C. Wallace

Just as using identical cameras is important for calibration, so too must the negatives be of identical size and resolution. While archival scans are produced on differing scanners at differing resolutions, they may also be cropped at different dimensions to remove elements that are superfluous to the subject. In order to produce accurate calibration data, scans of original negatives and new calibration negatives should be done on the same scanner, in the same environment, and at the same resolution. If this is not possible, the original scans should be scaled to the size and resolution of the calibration scans. As mentioned earlier, a number of significant aerial documentations of archaeological sites have been done using blimps mounted with a Hasselblad EL 500 with Carl Zeiss Distagon Lens1:4 F ¼ 50 Mm with a 51.3-mm focal length (Fig. 55.3). Subsequent naming variations of the camera had nothing to do with changes in optics or focal plane but rather the method of loading the film. A Hasselblad EL 500 was located, and Stephen Perry of Highart photography, New Zealand, took photographs of a calibration pattern (checkered screen). The resulting film was then scanned on an Epson V850 Pro scanner (13858x13858 pixels) alongside the original aerial photographic negatives taken of the Athenian Agora and Ancient Corinth by J. Wilson Myers and Eleanor Emlen Myers. The scanning took place in the same temperature and humidity so that at the high resolutions they were being scanned at, there would not be small differences in the size of the negatives due to expansion and contraction, ensuring the highest possible correlation between the old and the new. Fig. 55.3 Hasselblad EL 500 with Carl Zeiss Distagon Lens1:4 F ¼ 50 Mm (https:// commons.wikimedia.org/ wiki/File:Hasselblad_500_ CM.jpg)

55

Retrospective Photogrammetry: “Building a Time Machine”

1659

Table 55.1 Omega House model accuracy as affected by high res. Scanning and camera calibration Omega House models with contemporary survey Original varied photos from digital archive High res and archive photos High res only High res with calibration

Overall error in meters 3.61555 0.365096 0.446259 0.161563

Overall error (pixels) 12.522 1.717 6.915 2.870

The scanned calibration photographs were then processed through PhotoScan Lens to produce a calibration data file, which was then imported into PhotoScan and applied to the corresponding scans. It should be noted that although retrospective photogrammetry is intended to include both close range and aerial photographs, for the purposes of the calibration experiment, only the aerial photographs taken with the Hasselblad were used to ensure consistency. Once applied, the overall error in meters was reduced from 0.4462 to 0.161563 (Table 55.1). It should be noted that PhotoScan does not denote an error based on actual measurement vs model measurement but rather compares the model result to what would be expected based on the site survey. Therefore, any error in the survey is included in the model error. As can be seen in Figs. 55.4 and 55.5, the application of calibration has a significant effect in improving the distortions of the model produced by lens errors. Close range photographs of Omega House were taken by professional photographers using large format cameras and by archaeologist John Camp using a Leica M2 with a 50-mm Leitz Wetzlar 1:2 Lens. As with the Hasselblad, there was also a certain amount of standardization with the same model of camera being used on at multiple archaeological sites, and so an identical Leica was located in Ancient Corinth and James Herbst photographed calibration screens, which provided 35-mm negatives to produce calibrations for the close range SLR photographs taken of Omega House in the early 1970s (Fig. 55.6). The negatives were scanned at 6480x3999 pixels. Time limitations prevented the scanning of the original Leica negatives, but high-resolution scans were later made available and these were resized to 6480x3999 pixels as well. Because 35-mm photographs are rectangular and the photographs were either taken vertically or horizontally, two separate calibration files were produced. Table 55.2 includes each calibration separately as well as the result of both calibration applications combined. Experiments in modeling used the Leica SLR photographs exclusively, once again, in order to test calibration results without introducing other variables. Therefore, there is no survey data involved and thus no measurement of error in meters, but with the addition of calibration data, the overall error in pixels dropped from 0.55

1660

C. Wallace

Fig. 55.4 Omega House model without calibrated photographs

Fig. 55.5 Omega House model with calibrated photographs

to 0.062 (Table 55.2). As with the aerial modeling, a considerable visual improvement can be seen in Figs. 55.7 and 55.8. Wall geometries are significantly straightened and artifacts and distortions are removed.

55

Retrospective Photogrammetry: “Building a Time Machine”

1661

Fig. 55.6 Leica M2 with a 50-mm Leitz Wetzlar 1:2 Lens

Table 55.2 Omega House with gradual addition of calibration Omega House stairway Photos same size Stairs calibrated horizontally Stairs calibrated vertically Stairs calibrated vertically and horizontally Optimized stairs calibrated vertically and horizontally

55.4

Overall error (pixels) 0.550 0.425 0.210 0.062 0.056

Applications in Archaeological Documentation and Restoration

In recent years, some of the most important archaeological discoveries have been made, not in the field, but in museum basements and archives (https://www.rom.on. ca/en/about-us/newsroom/press-releases/massive-barosaurus-skeleton-discovered-

1662

C. Wallace

Fig. 55.7 Omega House stairway without calibration

Fig. 55.8 Omega House stairway with calibration

at-the-rom) (https://www.nytimes.com/2014/03/20/arts/artsspecial/golden-age-ofdiscovery-down-in-the-basements.html) (https://www.iol.co.za/news/missingancient-scroll-found-in-museum-basement-1282518). Photo archives hold the same promise. For decades archived photographs were seen as solitary, static representations of portions of sites, monuments, and artifacts. With the methods outlined above, these collections can yield new data and perspectives.

55

Retrospective Photogrammetry: “Building a Time Machine”

1663

While concerns have been raised about the data size of 3D models in archives [3], in actuality, a single, high-quality 3D model stored in an archive is actually the equivalent to an infinite number of photographs. Contemporary excavations are increasingly required to document in three dimensions, and so too older excavations can benefit from the inclusion of models that allow metric investigation and alternate views of sites and features. In Greece, foreign institutes have been excavating for over a century, and their predominant goal has been to gain access to Hellenic and Roman levels of stratigraphy, digging through, dismantling, and destroying newer layers of Byzantine, Frankish, Venetian, and Ottoman structures. In some cases, sufficient photographs and drawings have been preserved to enable their digital reconstruction. To date, this potential has been largely untapped. There are currently several projects aimed at addressing this. One of the most important contributions that retrospective photogrammetry is likely to make in the future is in the documentation and reconstruction of lost cultural heritage. Ongoing attempts by groups like Pavelka et al. [33] (in Syria) illustrate how the inspiration for the resurrection of lost cultural heritage has been found in examples like the early digital reconstructions of the Bamiyan Buddha in Afghanistan [20], and in an era where so much of our shared cultural heritage has been devastated by warfare and fanaticism, this technology and these efforts could not have arrived at a better time. The academic need to access previous archaeological information has been ongoing for over a century, and the efforts of those like Zawieska et al. [53] and myself [47] are in aid of creating resources that future academics will be able to use to apply innovations in archaeological interpretation while revisiting cultural resources that have changed or no longer exist. The ability to utilize photogrammetry in the restoration of damaged and deteriorated archaeological structures is going to become essential. For example, there is a desire in the Hellenic Ministry of Culture and Sports to conserve and restore Omega House in the Athenian Agora, which has significant water features, mosaics, and architecture. With that aim in mind, Omega House was modeled three dimensionally in its current state as well as being modeled using archival photographs taken when it was first excavated. The resulting models are being used to determine both the level of deterioration and the viability of restoration. In this case, with detailed modeling, we are able to examine individual walls and work with engineers to move forward in the preservation of this important historic site. In 2016, Discamps et al. published findings on the use of archival photographs of Middle Paleolithic excavation stratigraphy at Combe Grenal, Le Moustier, and Regourdou, France. They model very two-dimensional faces, and their goal is not as much to create 3D models as it is to georeference the archival records of the excavation. With this in mind, few photographs were necessary (two for Combe Grenal, three for Le Moustier, and no overlapping photographs were found for Regourdou). They found that despite the model’s lower resolution and poorer quality, when compared to contemporary models of the same locations, “their creation in parallel to fieldwork provides a critical and efficient tool for guiding decisions taken during archaeological excavations” [13].

1664

C. Wallace

At CAA, Tubingen 2018, Thomas Wolter presented similar work involving ongoing findings in the use of three-decade-old excavation stratigraphy photographs in combination with current excavation documentation to analyze stratigraphic faces that no longer exist in relation to those being currently explored. He stated in his abstract that “we are now trying to build a referenced Photogrammetry Model from 1980’s and 2015th campaign Photos and drawings” [52]. This is one of the directions that archaeological photogrammetry is heading, and so steps have been taken to photographically document various levels of archaeological excavations in Canada and Greece over the past several years in order to keep a record of their progress and enable comparative modeling in the future.

55.5

Error, Accuracy, and Uncertainty

“Measurement in the physical sciences is typically regarded as resulting in observed values composed of the true value plus an additional random component” [21]. The use of the terms “error” and “uncertainty” in discussion forums, journal articles, and software manuals often switches their meanings, one being used as the other. “Use of the term uncertainty seems to signal an appreciation of vagueness as well as randomness in geographical information, while error may send a misleading signal that true values are definable and retrievable” [54]. For unskilled users, retrospective photogrammetry can seem more like rolling dice, and indeed, it is a worthwhile step at the beginning of a modeling project to throw all of those photographs together with no guidance or parameters just to see how they associate with one another. This can determine next steps. However, ultimately the purpose of photogrammetry is that it needs to be quantifiable and measurable and its accuracy and uncertainties understood. To achieve this, we must first understand, within the context of the photogrammetric process, what these qualities mean. “Modelling represents a complex interaction of human and instrumental factors, and acquiring the raw components of each model, the data themselves, is also subject to a host of uncertainties. Depending on the skill of data analysts and the sophistication of instruments, data acquisition will have varying levels of error” [54]. All of these uncertainties are exacerbated in the context of retrospective photogrammetry, and so we must develop methods that reduce or negate them. Due to their temporal nature, contemporary photogrammetry can be measured more easily for error, while the error in retrospective photogrammetry, if we are objective about it, is more difficult to determine due to greater uncertainties. Even when we introduce georeferencing and other measurement data, that data is questionable and possibly unverifiable. In some cases, with limited or unreliable measureable references, photogrammetry taken from archival photographs is, at best, a visualization of what an archaeological site or artifact was, with dimensions that are approximate. Even without high accuracy, that 3D visualization can help in the restoration or understanding of the subject. Retrospective photogrammetry is more of a heuristic process wherein, while striving for the goal of accuracy and

55

Retrospective Photogrammetry: “Building a Time Machine”

1665

measurability some methods may seem irrational or counterintuitive, it is the result that is important. We can approximate the past, verify similar attributes, but never actually revisit it in its time to verify our results. For example, if an archaeological site still exists and seemingly stable architectural and geological features remain intact, we cannot guarantee that their current positions are exactly the same as they were 60 or 100 years ago. Manipulation by excavators, erosion of object surfaces, seismic activity (a particular problem in Greece and the reason why the Athenian Acropolis is constantly monitored by accelerographs and optical fibers [25]), settling due to underground water systems, and even rebound due to excavation can all contribute to variations in the actual position of a point, which results in measurement uncertainty and negates the ability to assess actual error. Even if we apply original survey data or take measurements from site plans, having no means of verifying them leaves us only with uncertainty. This is what [58] refer to as temporal uncertainty.

55.5.1 Measurement Error Measurement error is the difference between the values we measure and their “true” value [18]. For example, the error in calibration, positioning of a total station, or how its corresponding prism is positioned results in a difference between recorded distances and actual distances (true value). In the case of Omega House, a fourthcentury A.D. sprawling residential structure in the Athenian Agora, our survey derived from a datum at the north end of the Agora, while Omega House is at the south end; thus, the datum was far outside the perimeter of the area studied. In the past 50 years, there has been a significant growth of landscape trees in the Athenian Agora, which now causes problems with lines of sight for surveying. Consequently, surveying with the total station involved referencing a series of three markers external to the Omega House before surveying of the actual site could take place. No datums were located within Omega House. Each of these extra surveying steps introduced propagated measurement errors. In the survey of points on the site, the three most egregious errors occurred with the three points furthest from the total station. PhotoScan photogrammetric 3D modeling software shows error in meters for individual points and calculates an average. However, that error is actually the difference between the final position of the point within the model and how PhotoScan’s calculations expected that point to be placed. There is no comparison to a true value, and so it is in fact an estimation and not error. Measurement error would be the difference between the distances between points in the model and physically reality of the points on the site. In another example, if, in a 40 Celsius site, we accurately trace the tiles of a mosaic on mylar and then attempt to reassemble that mosaic in a 25 Celsius environment, the mylar will have contracted and we have a measurement error. All means of measuring have an inherent error which only varies by degree. With

1666

C. Wallace

knowledge of these errors as produced by different methods, we can minimize but in practice not eliminate them. We also have to address the problem of camera errors and how they contribute to measurement error. Dai et al. [12] conclude that measurement errors result from two large categories: (1) the systemic error due to camera factors and (2) the systemic error due to poor planning of camera network geometry. The latter is beyond the control of the retrospective photogrammetrist because it involves archival photographs, and so it is in the area of camera factors that we focus on minimizing measurement errors. As previously mentioned, “All camera lenses have a certain amount of distortion which increases toward the edges of the lens. Some current photogrammetry software includes catalogues of calibration corrections for known contemporary digital cameras. Metadata embedded in the photographs taken with these cameras identifies camera, lens, focal length, and settings at the time of exposure. Corrections are then applied to the photographs during modeling, reducing overall error.” The Agisoft software is able to make estimates of calibration corrections (when none are given) depending on the subject and quality of the photographs being used, but by producing calibration figures for the exact camera and lens combinations used in an archival collection, we can quantifiably reduce measurement error.

55.5.2 Scales Versus Surveying Three of the greatest advantages of close range photogrammetry over laser scanning and surveying are its cost savings, mobility, and flexibility. When surveying is not possible in contemporary photogrammetry, targets and scales can be used [42]. This does result in reduced measurement accuracy but can be remediated with properly used scales and/or targets. In the 3D modeling of the Eutychia mosaic in Ancient Corinth, using total station surveyed targets, we were able to achieve an accuracy of 0.4 mm [46], while Dai et al. [12], using unsurveyed scales and targets, allowed the software to determine scale and were able to achieve accuracies in the 3- to 4-mm range. In both cases, “ground truth” was measured for comparison rather than relying on the software’s assessment of accuracy. Therefore, those accuracy numbers are a measure of error and not uncertainty. In retrospective photogrammetry, it is highly improbable that a scale will be seen in one or more archival photographs, but if enough of a site is existent and surveying is unavailable, scales can be used to enhance the accuracy of the modeling. If we use ground scales and model the site in the present, we can produce points that can be used in the retrospective modeling. Even in a set of archival photographs that have no survey and no scales to be seen, there is still the possibility to create scale if any objects of a known dimension appear in the photographs. Standard sizes of boxes, shovel lengths, and even an inclusion of existent architectural features within the frames allow us to create points and attribute distances between them. While the use of scales on their own can help us to create a measurable model, they do not enhance the accuracy of the model in the way that a survey can.

55

Retrospective Photogrammetry: “Building a Time Machine”

1667

55.5.3 Measurement Uncertainty Uncertainty of measurement is the doubt that exists about the result of any measurement. You might think that well-made rulers, clocks and thermometers should be trustworthy, and give the right answers. But for every measurement – even the most careful – there is always a margin of doubt. In everyday speech, this might be expressed as ‘give or take’ . . . e.g. a stick might be two metres long ‘give or take a centimeter.’ [7]

Measurement uncertainty can be confused with measurement error. While error is the difference between a measured value and the true value of measurement, uncertainty is the range of possibilities for that measurement: a means of designating a range of values within which lies the “true” value of the measurement [27]. As stated by Couclelis [11], uncertainty occurs simply because part of the information is unknown or, as Fisher [17] puts it, it cannot be known with precision.“Kardos et al. [58] identify three types of uncertainty: attribute uncertainty, which applies to differences existing between the semantic characteristic of a feature and the corresponding data stored; spatial uncertainty, which is relative to differences between the actual physical position of a feature and the corresponding data stored; and temporal uncertainty, the time difference between data acquisition and data utilization” [45]. In an existent archaeological site, we can measure and compare accuracy of a model, but in a damaged or nonexistent archaeological site, we only have uncertainties. Agisoft PhotoScan has indicators of what it calls “metric error,” but that error is the difference between input reference points and what the software estimated for those coordinates. As such, in retrospective photogrammetry, what is actually being referred to is measurement uncertainty, and “the notion of uncertainty can be a more realistic view of . . . measurement than the notion of errors, which suggests attainable true values” [54]. Therefore, if the average “error” for all points is 1 meter, the model has, in fact, a measurement uncertainty of 1 meter. Relative measurement uncertainty takes into account the size of what is being measured. That relativity is important in that an uncertainty of 1 meter can be acceptable in modeling derived from aerial photographs and yet disastrous when pertaining to a model of an area only a few meters across. The only way we can cross into accuracy is if we compare the measurements of the model to physically measured site features.

55.5.4 Positional Error and Uncertainty Just as the surveyor, contemporary or in the past, arbitrarily chooses what points to survey, so too the person choosing reference points in structure from motion is choosing points to tie together the photographs in the models. While the software is also choosing points for its sparse point cloud, it does not have a familiarity or an understanding of the points that it is choosing. In the case where several nearly

1668

C. Wallace

identical architectural features might be misinterpreted by the software, the photogrammetrist, through familiarity with the photographs and the site, can discern differences during georeferencing. However, both positional error and positional uncertainty can be introduced. Although positional accuracy and positional uncertainty are conceptually different phenomena, it is very often difficult or impossible to differentiate between the two from an operational point of view—that is, to measure one separately from the other. This is especially the case in historical cartography, for the obvious reason that recourse to fieldwork is not an option. [45]

Positional accuracy can refer to the accuracy of survey measurements, manually inserted reference points, and the final destination that the software attributes to both. For example, if we physically measure the distance between two stable, existent points on an archaeological site at 40.29 meters and we measure those same two points on our retrospective model at 39.92 meters, there is an error of 0.37 meters, which translates to an error of 0.149%. If, however, we allow the software to estimate the “error,” it is actually an uncertainty. Positional uncertainty can involve the process of placing those surveyed and manually placed reference points and the vagueness in some photographs of those points due to focus, pixels, distance, etc., leading to a random estimation of where the actual point is by the photogrammetrist. “Positional accuracy can be separated into two classes, absolute and relative, where relative positional accuracy describes the consistency of any position on a map with respect to any other and absolute accuracy is a measure of deviation of an estimate from the true value” [8]. By saying so, Canavosio-Zuzelski et al. imply that error and uncertainty are the equivalent of absolute accuracy and relative accuracy. When working with archival photographs, there are visual uncertainties due to different camera lens distortions as well as image quality and distance. When placing a marker into a photographic point that is a meter from the camera, there can be an uncertainty of 1 cm, whereas a point in the distance can have an uncertainty of 1 meters.

55.5.5 Image Quality Error and Uncertainty “The accuracy of digital photogrammetry mainly depends on the camera-object distance, image quality and resolution; the latter is related [t]o the digital camera characteristics or to the scanner resolution” [15]. Their reference to scanner resolution is particularly important to retrospective photogrammetry. Poor input, e.g. vague photos can influence alignment results badly. To help you to exclude poorly focused images from processing. PhotoScan suggests automatic image quality estimation feature. Images with quality value of less than 0.5 units are recommended to be disabled and thus excluded from photogrammetric processing, providing that the rest of the photos cover the whole scene to be reconstructed. (Agisoft User Manual)

55

Retrospective Photogrammetry: “Building a Time Machine”

1669

As stated previously, focus, depth of field, object distance, scan quality, and image diversity (size and quality) all contribute to uncertainty. Error, however, is affected by lens distortion of the camera, the photo enlarger (when involved), and the scanner. The use of JPG files introduces error due to the compression of the image. Where possible, it is better to use a RAW photographic format. A general principle in archaeological photogrammetry is that a metric error of 2 cm is an acceptable goal [23, 39]. Of course, when including aerial photographs to tie the model together, their greater range reduces that accuracy overall but does allow higher localized accuracies. While there are no standard best practices for metric error in aerial photogrammetry, 5–20 cm seems to be repeated in drone standards [34]. In some aerial photographs, there is a record of altitude, and we can have an idea of the resolution that we are dealing with but with most there is no record. In the notebook records of the blimp recordings of Omega House, we have data documenting the altitude that the blimp was deployed at.

55.5.6 Error and Uncertainty in the SfM Process When producing a model through SfM (structure from motion), if we know the distortion of a particular camera, then we know the error of the images that it produces. With that knowledge, we can adjust our photographs to remove that distortion, particularly when it is a DSLR in our hands that we can calibrate, or use documented calibration data if it is available. When we do not know the camera and lenses that were used in an archival photo set, SfM produces errors caused by the distortion of the images that we are using. This is a major challenge of retrospective photogrammetry. Contemporary photogrammetry involves error that can be accounted for and, to a certain extent, can be corrected for. Retrospective photogrammetry involves significantly more error because in almost all cases, we lack the EXIF (Exchangeable Image File: the metadata stored in a digital image including camera and lens information, exposure, focal length, ISO, etc.) data from the camera or even knowledge of what camera was used. Different unknown cameras cause a propagation of uncertainty. In the absence of calibration data, Agisoft PhotoScan attempts to estimate the camera’s error, and documentation of the error correction is available in the software’s reporting function. In addition, we cannot verify the accuracy of any archival plans or drawings if their features do not exist today, and so a greater uncertainty prevails. Uncertainty and error in the SfM process can be quantified as shown by Russell [38]. Russell uses an orthophoto produced from the SfM model and compares measurements to the GPS positions for his targets to get x and y deviation. He uses a Dem produced by the SfM and does the same with the targets in order to get the z deviation. He then calculates root mean square error (RMSE) as follows:
0:5 d ¼ h2 þ e 2

ð55:1Þ

1670

C. Wallace

d is total offset distance. h is horizontal offset. e is elevation difference. rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pn 2 2 i¼0 d i RMSE ¼ n

ð55:2Þ

RMSE is the root-mean-square error. di is the total offset of each GCP. n is the number of GCPs used [38]. Once he had the RMSE of the model, he “propagated the uncertainty of the GCP measurements with the RMSE of the SfM model” [38] in order to determine uncertainty as compared to the WGS84 datum.
0:5 u ¼ RMSE2 þ h2 þ v2

ð55:3Þ

u is the uncertainty of the SfM model. RMSE is the root-mean-square error. h is the horizontal accuracy of the GCPs. v is the vertical accuracy of the GCPs [38].

55.5.7 Image Georeferencing Georeferencing, when possible, is a means to reduce both error and uncertainty; and yet both are present within the georeferencing itself. With georeferencing in the modeling, we are given both pixel error and metric error by the software. If the pixel error is low but the metric error is high, it can be an indication that there is an error in the coordinates we are using. In the case of Omega House, the overall error for the survey data from a Leica total station was 0.548528 with points 1, 6, and 8 producing the worst errors. After removing points 1, 6, and 8, the average metric error was 0.162517 meters. After modeling using the survey data, camera optimization, camera calibration, gradual filtering, and additional, unsurveyed ground control points (GCPs), the result was an overall metric error of only 0.119543 meters, a significant achievement when the average distance (altitude) from the subject was 36.3 meters with a ground resolution of 2.49 mm/pix.

55.5.8 Image Feature Detection Modeling of the sparse cloud does not involve finding common points but rather common features. At the outset of modeling, it is advantageous to manually deselect photographs that have uncertainties based on their lack of proximity to

55

Retrospective Photogrammetry: “Building a Time Machine”

1671

the object, focus, resolution, and shallow depth of field. This process will reduce the overall processing time for the model. Once done, the software then is able to search for feature matches. The settings for producing a sparse point cloud which rank from highest to lowest involve the software removing photographs with a higher level of uncertainty based on these criteria. If we process the photographs to correct for lens distortions prior to selection, we have addressed error reduction as well.

55.5.9 Image Reconstruction Error and Uncertainty When using PhotoScan, there is a refinement methodology used to correct both error and uncertainty called “Gradual Filter Settings.” Photogrammetrists using contemporary photographs generally use the default values in each step. However, in retrospective photogrammetry, it is a much more gradual process due to the inconsistencies between the photographs. In retrospective, it is advisable to ignore reprojection error, use 12 pixels as the reconstruction uncertainty and then work down to ten if possible (too many features might disappear), and set the projection accuracy to 4.5 pixels, the latter removing erroneous, out-of-focus, or otherwise egregious points. The following are the settings suggested by Agisoft as “default” settings. Parameter Reprojection error Reconstruction uncertainty Projection accuracy

Units Pixels Pixels

Value 0.2 10

Description Statistical error in point placement Possible variation in point placement

Pixels

10

Accuracy of point placement from local neighbor points

Agisoft forum [1]

These settings change radically when using older, uncalibrated photographs with no EXIF information. It is only through hundreds of retrospective models that the most beneficial settings, listed in the chart below, are arrived at. The Gradual Filter Settings do not work with retrospective photogrammetry due to the amount of uncertainty in the photographs. A value of 0.2 pixels for the reprojection error will make virtually all of the points disappear. Below are the settings that the author has found to give optimum results when working with archival photographs. These vary depending on the quality and similarity of the photographs used. Parameter Reprojection error Reconstruction uncertainty Projection accuracy

Units Pixels Pixels

Value 1.5 12

Description Statistical error in point placement Possible variation in point placement

Pixels

4.5

Accuracy of point placement from local neighbor points

1672

55.6

C. Wallace

Conclusions

With each project, it becomes easier, both objectively and subjectively, to determine which sets of archival photographs are able to be used to produce viable 3D models. However, the lessons learned within each project are also expanding that viability with photo sets that would once have been rejected now meriting a second look. This is due, in part, to having access to original archival negatives for scanning as well as the use of camera calibration data resulting in significant gains in accuracy. As more camera/lens combinations are determined within site contexts, those cameras will be sourced and calibration numbers documented. As a result, a greater number of archaeological sites will be able to be accurately modeled in 3D. In cases where there is the possibility through anecdotal information that a particular camera was used, we can apply those calibration corrections and see if there is the expected drop in error similar to that of known cameras. Fortuitously, some long-term excavations have retained many of the original cameras used, and with further interaction with the people originally involved and the retrieval of camera information from field notes and publications, a broader range of site photographs can be given the benefit of error reducing refinements. One can more easily assess and quantify error and uncertainty in contemporary photogrammetry and SfM. With retrospective photogrammetry, however, the vast number of unknowns, while being able to be defined as error or uncertainty, leave more aspects uncertain and unverifiable than not. Although major advancements have been made in SfM and digital photogrammetry, there is consistently a human element involved and nowhere more that in retrospective photogrammetry. A human perception of the elements being examined is essential for the success of the modeling. That being said, there are empirical elements of error and uncertainty that can be examined and honed in order to bring modeling of archaeological sites to their highest level. Quantifying temporal uncertainty is like quantifying the geometry of a cloud. If we are able to revisit sites for surveying and measurement, calibrate old cameras, and in general verify data, we can come closer to quantifying errors, but we depend on the veracity of field reports, memories of the original participants, and the stability of the original site features and so they remain uncertainties.

55.7

Moving Forward

As information about which cameras have been used in different excavations emerges, and calibrations of those cameras are acquired, there are a significant number of archaeological sites that can be reconstructed using all of the techniques mentioned in this chapter in order to produce quantitative retrospective photogrammetric modeling. Archives of the Myers photographs exist across Greece, site directors are still available to interview and find out which cameras were used on their excavations, and, surprisingly and significantly, many sites did not get rid of the old photographic equipment that was used in periods dating back 50, 60, and

55

Retrospective Photogrammetry: “Building a Time Machine”

1673

90 years. The potential for applying calibration to archival photographs is overwhelming, but the promise of results is worth the effort. More experimentation is required with varied sites with the current calibration data and photogrammetric techniques. Tests of software are required in order to determine why they do or do not succeed in producing retrospective models. Identification of cameras used and sourcing either the originals or exact matches will further calibration and the accuracy of our models. Most importantly, the people who were involved with excavations need to be interviewed concerning which cameras were used for which photographs while they are available because much of this knowledge was not recorded. Further research into archaeological field note books is also warranted in order to determine cameras used. With the calibration data in hand, it is possible to experiment with other photogrammetric software programs that do not allow the use of uncalibrated photographs and compare results. There is a real risk, however, that as photogrammetric software advances, it may become more dependent upon metadata and databases focused entirely on digital photography. What retrospective photogrammetry is becoming is a time machine through which archaeologists, researchers, and civilians can examine and experience places that no longer exist, not through digital recreations but as they actually looked when they were documented. This is a vital component in the preservation of our shared world heritage. Acknowledgments Special thanks go to Stephen Perry of Highart photography, New Zealand; James Herbst, John Camp, Craig Mauzy, and Bruce Hartzler of the American School of Classical Studies at Athens; and Betsey Robinson of Vanderbilt University, Tennessee, for all of their kind assistance.

References 1. Agisoft Forum (2017). https://www.agisoft.com/forum/index.php?topic¼6888.0 2. Agisoft PhotoScan User manual – professional edition, version 0.9 3. Agrawal R, Nyamful C (2016) Challenges of big data storage and management. Glob J Inf Technol 6. https://doi.org/10.18844/gjit.v6i1.383 4. Albertz J (2001) Albrecht Meydenbauer- Pioneer of photogrammetric documentation of the cultural heritage, Proceedings 18th international symposium CIPA 2001, 19 5. Baldi P, Cenni N, Fabris M, Zanutta A (2008) Kinematics of a landslide derived from archival photogrammetry and GPS data. Geomorphology 102:435–444 6. Barazzetti L, Mussio L, Remondino F, Scaioni M (2011) Targetless camera calibration. Int Arch Photogramm Remote Sens Spatial Inf Sci 38:Part 5/W16, 8 pages 7. Bell S (1999) A Beginner’s guide to uncertainty of measurement. Measurement Good Practice Guide. II 8. Canavosio-Zuzelski R, Agouris P, Doucette P (2013) A photogrammetric approach for assessing positional accuracy of OpenStreetMap© roads. ISPRS Int J Geo-Inf 2:276–301 9. Cantoro G (2015) Aerial_photogrammetry. When archaeology meets SIFT. In: Cowley D, Ivanišević V, Veljanovski T, Kiarszys G, Bugarski I (eds) Recovering lost landscapes. Institute of Archaeology, Belgrade

1674

C. Wallace

10. Chandler JH, Cooper M (1988) Monitoring the development of landslides using archival photography and analytical photogrammetry. Land Mineral Surv 6(11):576–584 11. Couclelis H (2003) The certainty of uncertainty: GIS and the limits of geographic knowledge. Trans GIS 7:165–175 12. Dai F, Feng Y, Hough R (2014) Photogrammetric error sources and impacts on modeling and surveying in construction engineering applications. Vis Eng 2:2 13. Discamps E, Muth X, Gravina B, Lacrampe-Cuyaubère F, Chadelle JP, Faivre JP, Maureille B (2016) Photogrammetry as a tool for integrating archival data in archaeological fieldwork: examples from the middle palaeolithic sites of Combe-Grenal, Le Moustier, and Regourdou. J Archaeol Sci Rep 8:268–276 14. Elefhteriou V, Manidaki V (2010) Photogrammetric documentation as a tool for architectural research in the acropolis works. Acropolis Restoration News 10:35–38 15. Fabris M, Pesci A (2005) Automated DEM extraction in digital aerial photogrammetry: precisions and validation for mass movement monitoring. Ann Geophys 48, 6 16. Falkingham PL, Bates JT, Farlow JO (2014) Historical photogrammetry: bird’s Paluxy River dinosaur chase sequence digitally reconstructed as it was prior to excavation 70 years ago. PLoS One 9(4):e93247. https://doi.org/10.1371/journal.pone.0093247 17. Fisher PF (1999) Models of uncertainty in spatial data. In: Longley P, Goodchild MF, Maguire D, Rhind D (eds) Geographical information systems: principles, techniques, management and applications. Wiley, New York, pp 191–205 18. Glen S (2016) Measurement error (Observational error). https://www.statisticshowto. datasciencecentral.com/measurement-error/ 19. Gruen A, Remondino F, Zhang L (2002) Reconstruction of the great Buddha of Bamiyan, Afghanistan. Int Arch Photogramm Remote Sens 34(5):363–368 20. Gruen A, Remondino F, Zhang L (2004) Photogrammetric reconstruction of the great Buddha of Bamiyan, Afghanistan. Photogramm Rec 19(107):177–199 21. Hand DJ (2004) Measurement theory and practice: the world through quantification. Hodder Arnold, London 22. Hemmleb M (1999) Digital rectification of historical images. In: International archives of photogrammetry and remote sensing, vol XXXII. CIPA International Symposium, 3–6 October 1999, Olinda 23. Howland MD, Kuester F, Levy TE (2014) Photogrammetry in the field: documenting, recording, and presenting archaeology. Mediter Archaeol Archaeom 14(4):101–108. Retrieved from https://escholarship.org/uc/item/5ps0z7pf 24. Johnson GW, Kase EW (1977) Mapping an ancient trade route with balloon photography. Photogramm Eng Remote Sens:1489–1493 25. Kapogianni E, Kalogeras I, Psarropoulos P, Melis N, Eleftheriou V, Sakellariou M (2017) Monitoring of the hill and the circuit wall of the athenian acropolis utilizing optical fibre sensors and accelerographs. Glob J Res Eng 17(2). Version 1.0, Publisher: Global Journals Inc. (USA) Online ISSN: 2249-4596 & Print ISSN: 0975-5861 26. Kolbl OR (1976) Metric or non-metric cameras. Photogramm Eng Remote Sens 42(1):103–113 27. MacKay B (2019) What is Measurement and Uncertainty?, from “Mathematics and Statistics Models”. https://serc.carleton.edu/sp/library/uncertainty/what.html 28. Maiwald F, Vietze T, Schneider D, Henze F, Münster S, Niebling F (2017) Photogrammetric analysis of historical image repositories for virtual reconstruction in the field of digital humanities. In: The international archives of the photogrammetry, remote sensing and spatial information sciences, volume XLII-2/W3. Proceedings of the 3D virtual reconstruction and visualization of complex architectures. 1–3 March 2017, Nafplio 29. Mavromati D (2010) Orthophotography and monuments. The problems of photogrammetric documentation of the ancient monuments. Acropolis Restoration News 10:16–19 30. Micheletti N, Chandler JH, Lane SN (2015) Structure from Motion (SfM) photogrammetry. In: Cook SJ, Clarke LE, Nield JM (eds) Geomorphological techniques. British Society for Geomorphology, London, pp 1–12

55

Retrospective Photogrammetry: “Building a Time Machine”

1675

31. Moullou D (2010) The project “Development of geographical information systems on the acropolis of Athens”. Managing scientific questions of documentation. Acropolis Restoration News 10:12–15 32. Moullou D, Mavromati D (2007) Topographic and photogrammetric recording of the acropolis of Athens. Proceedings of XXI international CIPA symposium. 1–6 October, Athens, pp 515–520 33. Pavelka K, Šedina J, Raeva P, Hůlková M (2017) Modern processing capabilities of analog data from documentation of the great Omayyad mosque in Aleppo, Syria, damaged in civil war. Int Arch Photogramm Remote Sens XLII-2(W5):561–565. https://doi.org/10.5194/isprs-archivesXLII-2-W5-561-2017 34. Pepe M, Fregonese L, Scaioni M (2018) Planning airborne photogrammetry and remote-sensing missions with modern platforms and sensors. Eur J Remote Sens 51(1):412–436. https://doi.org/ 10.1080/22797254.2018.1444945 35. Peppa MV, Mills JP, Fieber KD, Haynes I, Turner S, Turner A, Douglas M, Bryan PG (2018) Archaeological feature detection from archive aerial photography with a SfM-MVS image enhancement pipeline. Int Arch Photogramm Remote Sens Spat Inf Sci 42(2):869–875. ISPRS technical commission II symposium 2018 “toward photogrammetry 2020”, 3–7 June 2018, Riva del Garda, 869–875 36. Powlesland D (2014) 3Di – enhancing the record, extending the returns, 3D imaging from free range photography and its application during excavation. In: Kamermans H et al (eds) The three dimensions of archaeology (Proceedings of the XVII UISPP world congress, 1–7 September 2014, Burgos, Spain), vol 7. Archaeopress, Oxford. 13–32. Received October 2016; revised September 2017; accepted November 2017 37. Remondino F, Menna F (2008) Image-based surface measurement for close-range heritage documentation. Int Arch Photogramm Remote Sens Spat Inf Sci 37(B5):199–2006 38. Russell TS (2016) Calculating the uncertainty of a structure from motion (SfM) model, Cadman quarry, Monroe, Washington, A report prepared in partial fulfillment of the requirements for the degree of Master of Sciences Earth and Space Sciences: Applied Geosciences University of Washington 39. Sapirstein P (2016) Accurate measurement with photogrammetry at large sites. J Archaeol Sci 66:137–145. https://doi.org/10.1016/j.jas.2016.01.002 40. Schindler G, Dellaert F (2012) 4D Cities: analyzing, visualizing, and interacting with historical urban photo collections. J Multimed 7(2) 41. Schultz G, Dallas RW (1985) 100 years of historical monument archives. Photogramm Rec 11: 761–763. https://doi.org/10.1111/j.1477-9730.1985.tb01329.x 42. Shervais K, Dietrich J (2016) Structure from motion (SfM) photogrammetry data exploration and processing manual. https://kb.unavco.org/kb/file.php?id¼781 43. Snavely N, Seitz SM, Szeliski R (2007) Modeling the world from internet photo collections. Int J Computer Vision November 2008 80(2):189–210. https://doi.org/10.1007/s11263-007-0107-3 44. Thompson EH (1962) Photogrammetry in the restoration of castle Howard. Photogramm Rec 4: 94–119. https://doi.org/10.1111/j.1477-9730.tb00334.x 45. Tucci M, Giordano A (2011) Positional accuracy, positional uncertainty, and feature change detection in historical maps: results of an experiment. Computers, Environment and Urban Systems 46. Wallace C (2016) Photogrammetry in Mediterranean archaeology. UWSpace. http://hdl.handle. net/10012/10344 47. Wallace C (2017) Retrospective photogrammetry in Greek archaeology. Stud Digit Herit 1(2): 607–626. https://doi.org/10.14434/sdh.v1i2.23251 48. Whittlesey JH, Myers JW (1975) Whittlesey foundation field season, 1975. J Field Archaeol 2(4):351–361. https://doi.org/10.1179/009346975791490908 49. Whittlesey JH, Myers JW, Allen CC (1977) The Whittlesey foundation 1976 field season. J Field Archaeol 4(2):181–196. https://doi.org/10.1179/jfa.1977.4.2.181 50. Wiedemann A, Hemmleb M, Albertz J (2000) Reconstruction of historical buildings based on images from Meydenbauer archives. Int Arch Photogramm Remote Sens XXXIII:Part B5, Amsterdam

1676

C. Wallace

51. Wilson T, Miles HC, Labrosse F, Tiddeman B, Roberts JC (2016) Historical records, archives and photogrammetry. Hist Environ 7(1):25–42. https://doi.org/10.1080/17567505.2016. 1142696 52. Wolter T (2018) Reconstructing excavations? Photogrammetry on 1980’s photos. Abstracts CAATuebingen 2018:60 53. Zawieska D, Markiewicz JS, Kopiasz J, Tazbir J, Tobiasz A (2017) 3D modelling of the Lusatian borough in Biskupin using archival data. Int Arch Photogramm Remote Sens Spat Inf Sci XLII-2/W3:665–669. https://doi.org/10.5194/isprs-archives-XLII-2-W3-665-2017 54. Zhang J, Goodchild MF (2002) Uncertainty in geographical information. CRC Press, ISBN 9780415243346 – CAT# TF1222 55. Adams J, Chandler J (2002) Evaluation of Lidar and Medium Scale Photogrammetry for Detecting Soft-Cliff Coastal Change. The Photogrammetric Record 17: 405–418. https://doi. org/10.1111/0031-868X.00195 56. Chandler J, Buckley S (2016) Structure from motion (SFM) photogrammetry vs terrestrial laser scanning (Version 1). Loughborough University. https://hdl.handle.net/2134/21673 57. Condorelli F, Rinaudo F (2018) Cultural Heritage Reconstruction from Historical Photographs and Videos, The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLII-2, 259–265. https://doi.org/10.5194/isprs-archives-XLII-2-259-2018 58. Kardos J, Moore A, Benwell G (2005) Exploring the Hidden Potential of Common Spatial Data Models to Visualize Uncertainty, Cartography and Geographic Information Science 32(4):359– 367. https://doi.org/10.1559/152304005775194827 59. Miranda ÁR, Melón JM (2017) Recovering old stereoscopic negatives and producing digital 3D models of former appearances of historic buildings. ISPRS - International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 601–608 60. Opitz R, Limp W (2015) Recent Developments in High-Density Survey and Measurement (HDSM) for Archaeology: Implications for Practice and Theory. Annual Review of Anthropology 44(2015):347–64 61. Walstra J, Chandler J, Dixon N, Dijkstra T (2004) Time for change: quantifying landslide evolution using historical aerial photographs and modern photogrammetric methods. INTERNATIONAL ARCHIVES OF PHOTOGRAMMETRY, REMOTE SENSING AND SPATIAL INFORMATION SCIENCES (Vol. 35, pp. 475–480). Presented at the 20th ISPRS Congress, International Society for Photogrammetry and Remote Sensing (ISPRS). Commission IV. Working Group 8

Part VI Conservation and Restoration

Operational Modal Analysis Method for Historic Masonry Structures: Applications

56

Dilek Okuyucu

Contents 56.1 56.2

Engineering Touch to Historic Masonry Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Creation of Structural Analysis Model of a Historic Masonry Building . . . . . . . . . 56.2.1 Masonry Modeling Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.2 Material Characterization for Macro Modeling . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.3 Verification of the Structural Analysis Model . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3 Operational Modal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.1 Instrumentation and Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.2 Estimation of Experimental Modal Parameters . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.3 Validation of Mod Shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4 Possible Applications of OMA for Historic Masonry Structures . . . . . . . . . . . . . . . . 56.4.1 Structural Model Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.2 Structural Health Monitoring and Damage Detection . . . . . . . . . . . . . . . . . . 56.4.3 Some OMA Application Examples from Author’s Research . . . . . . . . . . . 56.5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1680 1683 1684 1685 1686 1687 1687 1688 1689 1689 1689 1690 1690 1696 1697

Abstract

The chapter presents information about applications of operational modal analysis on historic masonry structures. Functional modal analysis is a nondestructive type of in situ experimentation technique that enables the estimation of the construction’s modal behavior parameters. Considering technical advantages, the method finds a wide range of applications in different fields of engineering. Operational modal analysis is preferred for some kinds of structural engineering studies on historic masonry constructions as well. In the case of historic masonry, the method can be preferred for structural analysis model calibration, verification of masonry material identification procedures, structural health monitoring, condition assessment, and damage detection. The text provides information about D. Okuyucu (*) Civil Engineering Department, Erzurum Technical University, Erzurum, Turkey e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_56

1679

1680

D. Okuyucu

masonry modeling approaches that link operational modal analysis to the macro modeling method. Possible operational modal analysis applications are stated in the manuscript; operational modal analysis studies of the author on miscellaneous historic masonry constructions are exampled for different purposes. Keywords

Architectural heritage · Historic masonry · Structural engineering · Macro modeling · Operational modal analysis Abbreviations

EFDD Emasonry fc FDD fk MAC OMA Rv

56.1

Enhanced frequency domain decomposition Elastic modulus of masonry, in MPa The compressive strength of natural stone defined by the use of concrete Schmidt Hammer test, in MPa Frequency domain decomposition The compressive strength of composite masonry, in MPa Modal assurance criteria Operational modal analysis Average of rebound values measured by horizontal application on natural stone measured by concrete Schmidt Hammer

Engineering Touch to Historic Masonry Structures

Historic masonry structures deserve special attention from structural and earthquake engineering points of view. They adopt traces of regional seismicity, provide information about local construction practices, and resemble the existence of past civilizations. They can show the details of successful masonry construction, like Erzincan Değirmenliköy Church did [1]. Erzincan city of Turkey experienced two major earthquakes in 1939 and 1992; almost no historic masonry structure withstood in the city center [2–4]. However, there was a survivor of both destructive ground motions at the vicinity of Erzincan called Değirmenliköy Church, shown in Fig. 56.1. It is a

Fig. 56.1 Views of Erzincan Değirmenliköy Chuch in Erzincan, Turkey

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1681

remarkable example of lightweight masonry material utilization in historic masonry construction, which may be the reason to withstand [5]. Historic masonry may provide valuable technical information about past earthquakes like another example from Turkey, the Double Minaret Madrasah. In Islamic architecture, madrasah is defined as an educational facility consisting of both educational units and residential parts for educators and students [6]. The Double Minaret Madrasah is almost eight centuries old, the late Seljukian period monument in the Turkish city of Erzurum. It is the biggest Seljukian madrasah on the Anatolian land, with an open-air atrium. A minaret is a unique tower of Islamic architecture constructed adjacent to or inside the mosques as well as madrasah buildings. Their primary function is to call Muslims to prayer while providing a visual focal point. In case of a prayer call, the caller appears on the minaret balcony and loudly calls. It is not possible to construct the minarets without balconies in Islamic architecture. However, the minaret balconies of Double Minaret Madrasah are absent in the meantime. Figure 56.2 provides some old gravure drawings and the current situation of the minarets of the mentioned madrasah. Although Deyrolle figured the minarets in 1869 without balconies as they currently stand, the minaret balconies are visible and existing in Godfrey’s drawings in 1833 and Texier in 1840 [7]. The minarets initially had the balconies, and then the balconies disappeared, destroyed, or damaged. The technical reason for the disappearance of the balconies may be found in historical earthquake reports of Erzurum. On October 27, 1843, an earthquake occurred in Erzurum, and an English representative called Mr. Curzon, who was in Erzurum for Turkish-Iran border negotiations, reports about the earthquake [8]. The report tells about the total collapse of some mosque minarets, the devastation of old city walls, and heavy damage to the Double Minaret Madrasah’s minarets. Besides, another strong ground motion is stated in historical documents that happened on July 24, 1852 [8]. This earthquake was also reported to cause a total collapse of lots of masonry residential buildings and heavy damage to some mosque minarets. Collecting all information together, the 1843 and 1852 earthquakes may be stated to be the reason for Double Minaret Madrasah’s minaret

Fig. 56.2 Old drawings and current situation of the minarets of Double Minarets Madrasah [7]

1682

D. Okuyucu

balcony disappearance. This information gathered from a historic construction can be valued for synthetic ground motion modeling, creating of a site-specific design spectrum, etc. Comparing to the history of humankind, architectural and constructional activities may be stated as relatively recent activities. The history of architecture is to be started almost 10,000 years ago [9]. Architectural design is related to the earliest civilization, and, at the same time, masonry appears as a technique for construction. Therefore, human evolution can be related to the history of architecture and the history of construction materials [10]. Historic masonry structures also deserve respect and attention from this point of view. No matter where they are located, historic masonry buildings are common values of humanity. They are not the inheritance from the ancestors, but they are consignations of the next generations. Therefore, they must be technically protected and carried to the future, without forcing them to bring back their origin on the timeline. This provides a significant risk of total demolishing of the architectural heritage. In this phase, structural engineering takes the scene and starts to touch the historic masonry structure technically. In the case of historic masonry protection and moving it to the future, the most commonly followed path is to collect information about the construction, evaluate the current situation of the structure, and then study the possible reasons that can threaten the building. Starting from the last task, accidental actions like earthquakes appear to be a significant factor that resulted in the total collapse of numbers of historic masonry constructions. Furthermore, these structures are not only influenced by extraordinary events like earthquakes. High-risk architectural heritage factors are fatigue and strength degradation, accumulated damage due to traffic, wind loads and temperature, soil settlements, high level of sound, and lack of structural understanding of the original builders [11–13]. Information gathering and condition assessment of historic masonry structures contain some challenging issues like: • Missing data about the geometry of the structure • Lack of information about the inner core structure of the load-bearing elements • Difficulty and expensiveness of characterization of nonhomogenous masonry material • Significant variation of mechanical properties of the masonry due to use of natural materials and workmanship • Unknown construction sequence • Significant changes in the core structure and constitution of structural elements associated with long construction periods • Unknown exiting, non-visible damage of the structure • Non-applicable regulations and codes [11] Considering all of these unknowns, challenges, and risks, a structural engineer takes the responsibility of architectural heritage and uses all possible tools to

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1683

understand, evaluate, rehabilitate, restore, and strengthen the historic masonry building. The latter may be stated as one of the most difficult engineering services of a structural engineer. During this activity, structural analysis models appear as important tools to evaluate the overall structural behavior. At the first stage, the engineer should understand, evaluate, and feel the historic masonry structure by making in situ site studies. Following, the structural modeling phase may start. Both for the nonhomogeneous characteristics of the masonry and the many unknowns about the structure, many hypotheses for creating the model of structural analysis follow as a result. It should be underlined that the model is not that of the real structure; it is only a representative. Therefore, the creation of a structural analysis model contains errors due to some enforced assumptions. It is suggested to use the structural model to analyze the masonry structure for easier overall evaluation and engineering intuition. At this phase, a question immediately arises to find the case’s answer, whether it is possible to create more reliable structural analysis models of historic masonry. Several geometrical and mechanical property assumptions are executed to create a finite element model of the historic masonry because of the nonhomogenous feature of masonry and many unknowns about the structure’s construction practice. Pioneer engineers, distinguished researchers, and dedicated colleagues have been spending significant effort to provide acceptable and reasonable answers to the question for the decades, so does the author of the herein manuscript. As to give an answer, the author studies the possible use of an operational modal analysis (OMA) method for in situ structural evaluation, damage detection, and structural model calibration for historic masonry buildings [1, 14, 15]. OMA technique enables to define in situ modal behavior parameters. The use of this field experimentation data in historic masonry structures is the main content of the chapter. The text provides general information about masonry modeling approaches, OMA method, and possible OMA technique applications for historic masonry structures.

56.2

Creation of Structural Analysis Model of a Historic Masonry Building

Masonry is a nonhomogenous material that consists of a masonry unit, mortar, and mortar-unit interface. Adobes, bricks, stone blocks, and others are common masonry units. Mud, clay, glue, bitumen, lime, or cement-based binding materials can be used as mortars. The enormous number of possible combinations generated by masonry unit geometry, nature, arrangement, and mortar characteristics increase doubts about masonry modeling accuracy. The various types of masonry’s mechanical behavior usually have a common feature: an insufficient tensile strength [11]. Among so many unknowns, historic masonry modeling should be started by adopting a modeling strategy depending on the simplicity and the accuracy desired.

1684

D. Okuyucu

56.2.1 Masonry Modeling Strategies Masonry can be modeled either as composite or micro modeling of individual elements (unit, mortar, interface) that may be adopted to make a more detailed numerical representation [16, 17]. Figure 56.3 depicts the general modeling strategies of masonry construction. Detailed micro modeling: The primary objective of micro modeling is to represent masonry from each component’s technical knowledge closely. Mortar, unit, and mortar-unit interface are all introduced to the model. Discontinuum elements simulate a mortar-unit interface; continuum elements separately represent unit and mortar. Mechanical properties like Poisson’s ratio, Young modulus, and, if required, inelastic material properties of masonry unit and mortar are all taken into account. The most challenging case is the numerical representation of the interface. The interface is a potential crack/slip plane with initial dummy stiffness to avoid continuum interpenetration that enables the unit’s combined action, mortar, and interface. This method is ideal for small structural elements with a special interest in extremely heterogeneous stress and strain conditions. Laboratory tests are required to obtain mechanical properties of all aspects experimentally. Modeling of interface deserves special attention [18, 19]. Simplified micro modeling: As the name implies, this approach is a simplified version of detailed micro modeling. The simplification is realized at the joints. Each joint is represented by a new element that consists of mortar and its neighboring two interfaces, as presented in Fig. 56.3. Hence, masonry units and simplified joints are numerically simulated by continuum and discontinuum elements, respectively. The masonry units are also expanded to keep the geometry unchanged. Since the Poisson effects of the mortar are neglected, the calculation accuracy is lowered. The masonry is no longer represented by elastic blocks connected by potential fracture/slip lines at the joints. Macro modeling: In this approach, masonry is accepted as a homogenous anisotropic continuum, meaning that whole masonry material is represented by a relationship that is established between average masonry stresses and strains that mainly requires masonry tests carried out on sufficiently large-scale masonry units. Macro models are applicable when the structure consists of solid walls with sufficiently large dimensions to ensure that the stresses are essentially uniform across or along a

Fig. 56.3 Modeling strategies of masonry construction. (a) Detailed micro modeling; (b) simplified micro modeling; (c) macro modeling [11]

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1685

macro length [11]. The macro modeling approach is generally adopted for practiceoriented engineering applications to evaluate the overall structural behavior. Each modeling approach finds a suitable field of application. One strategy cannot be preferred to any of the others. Macro modeling approach provides less timeconsuming analysis and the need for smaller computer memory compared to that of micro modeling. The author studies large-scale historic structures and adopts macro modeling approach. Therefore, the determination of the mechanical properties of the composite masonry is an area of interest.

56.2.2 Material Characterization for Macro Modeling Current building codes provide formulations to estimate masonry’s mechanical properties as a composite using mortar and masonry unit properties [20, 21]. For instance, the current Turkish Seismic Design Code, 2018, provides Eq. 56.1 for the masonry’s elastic modulus, where fk is the compressive strength of the composite masonry [20]. Besides, code first rules to do experiments on masonry units to obtain the compressive strength. If this is not possible, individual tests on mortar and masonry units are stated to be done. At the final worst condition of experimentation, masonry units’ mechanical properties must be experimentally obtained at least. Emasonry ¼ 750 f k

ð56:1Þ

It should be underlined that current building codes are applicable for new masonry constructions and they generally present rules for the masonry units like bricks and cut stones. The use of current building code regulations, like Eq. 56.1, to estimate the material properties of historic masonry was experienced by the author ending up with unacceptable technical results. The situation is completely different for existing historic masonry structures due to several unknowns, so the modern building codes are non-applicable to historic masonry. At this stage, engineers and researchers prefer to use nondestructive in situ experimentation for more reliable material characterization of historic masonry. Ultrasound velocity tests and rebound hammer applications are some of the in situ experiments [11]. Figure 56.4 shows a general view of the mentioned applications by the photos [22]. Ultrasound Velocity Tests: Ultrasound pulse velocity is a nondestructive testing method to obtain information about the surface coating materials’ mechanical characteristics and degradation state. It is based on the transportation of waves by vibration through the solid particles of a material. The more compact materials have a higher pulse velocity of propagation. Materials with weaker cohesion or the ones in poor conditions provide lower pulse velocity compared to that of compact and less deteriorated materials [23]. Thus, the ultrasound pulse velocity test allows the estimation of mechanical properties like internal mechanical strength or deformation capacity and the detection of discontinuities, such as cracks. The technique finds a wide range of applications on historic masonry structures, although it is a relatively expensive method [24, 25, 26].

1686

D. Okuyucu

Fig. 56.4 Site applications for material characterization of historic masonry [22]. (a) Ultrasound velocity test, (b) Schmidt hammer test

Rebound Hammer Tests: A Schmidt hammer, also known as a rebound hammer or concrete hammer test, is a device to measure the elastic properties or strength of rock or concrete, mainly penetration resistance and surface hardness. The method is generally known as the Schmidt hammer technique, and it depends on the determination of the concrete/rock compressive strength through surface hardness. Commercially, it is possible to find a rebound hammer for concrete and rock, separately. However, a concrete rebound hammer is more common and cheaper than that of the rock rebound hammer. Therefore, researchers and engineers tend to use concrete rebound hammer to determine the mechanical properties of masonry [27, 28, 29]. A concrete rebound hammer uses the calibration equations to convert surface hardness into concrete compressive strength. The practitioner should consider this consideration; surface hardness is obtained over masonry unit or mortar, and calibration sheets for concrete are used for an estimation! Studies that provide specific calibration equations that can be applied to the masonry using concrete rebound hammer, like reference [30], are needed.

56.2.3 Verification of the Structural Analysis Model Practice-oriented studies on historic masonry structures like condition assessment and dynamic behaviour evaluation are carried out very commonly [31, 32, 33]. The macro modelling approach is primarily adopted and material properties obtained through non-destructive experimentation are assigned. Yet, the model is created and ready for analysis? Not, sure. What is created on the computer is not the real structure; even it is suspicious to represent the real structure at an acceptable level of accuracy. The model contains numbers of unknowns, uncertainties and assumptions; the model needs verification. But, how? Can dynamic characteristics like mod shapes and related frequencies be used for model verification? By this way, the author can suggest a technical tool: operational modal analysis.

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

56.3

1687

Operational Modal Analysis

Operational modal analysis (OMA) is also called ambient response analysis, ambient modal analysis, output-only modal analysis, natural input modal analysis, and in-operation modal analysis. This is in situ experimentation to obtain dynamic behavior parameters (mode shapes, modal frequencies, and damping) of the structure. A typical modal test of a structure is performed by measuring the input forces and output responses for a time-invariant and linear mechanical system. The excitation is either transient, random, burst-random, or sinusoidal. The advanced signal processing tools preferred in OMA techniques allow the dynamic behavior properties of a mechanical structure (mode shapes, resonance frequencies, and damping ratios) to be estimated by only measuring the structure’s response without using an artificial excitation. This technique has been widely used in civil engineering structures (towers, buildings, platforms, bridges) where the natural excitation created by any source is used to extract modal parameters [34–36]. It finds more applications in mechanical and aerospace engineering (on-road testing, rotating machinery, in-flight testing) [37–39]. Moreover, OMA is also applied to historic masonry structures [40–43]. The advantage of this technique is that a modal model can be generated while the structure is functional. That is a model within true boundary conditions and actual force and vibration levels. Another advantage of the technique is the ability to perform modal testing in situ. OMA application can be performed with other applications or activities in parallel and does not affect the structure’s daily use. OMA application starts with output measurements from the structure and finalizes by estimation of in situ modal behavior parameters. An OMA test can be carried out following the main steps below: • Selection of measurement strategy depending on the number of available sensors and dimensions of the structure (simultaneous measurement or multi set-up) • Selection of degree of freedoms to be measured, and, if multi set-up (reference sensor-roving sensor method) measurement is decided, selection of reference sensor positions) • Design of the experiment (decision for optimal measurement parameters like frequency, total time, etc.) • Sensor instrumentation and data acquisition • Analysis of data quality (signal levels, stationarity, frequency content, etc.) • Modal parameter estimation (determination of modal parameters, validation of mod shapes by modal assurance criteria, system identification) The method’s advantages make OMA a preferred experimental method for dynamic identification of historic masonry for different purposes. Some important issues for OMA application to a historic masonry are stated in the following part.

56.3.1 Instrumentation and Measurements Ambient vibration measurements are taken from the historic masonry, which is an important advantage of OMA. High sensitivity accelerometers are preferred for

1688

D. Okuyucu

ambient vibration measurements and recordings. A sufficient measurement scheme and strategy (simultaneous or multi set-up) can be followed considering theoretical modal analysis results, and the mod shapes can be detected. The OMA application is essential to measure the structure’s responses to random vibrations and estimate the modal behavior parameters by processing this data. Therefore, it is extremely important to obtain vibration records from which the modal behavior parameters can be obtained from the structural response (measured acceleration records). As a common case in Turkey, in the area where the building is located, the most important vibrations reacted by the structure are generally the traffic-induced ground vibrations. Considering this, the measurements are preferred to be taken when the vehicle traffic is quite intensive. Vibration recording time depends on the spectral shape and duration of the vibrating signal, the presence of harmonic vibrations, the complexity of the structure under test, the quality of measurement equipment, etc. However, it is practically recommended to take measurements with more than 500 times the period value of the lowest mode shape taken into account, as a rough approximation. Besides, if the predominant harmonic vibrations in the measured responses are present, the recording time should be kept even longer [44]. Starting from this point, to carry out the theoretical modal analysis of the structure before the fieldwork is of primary importance. Theoretical calculation of mod shapes and related modal frequencies may help design a reliable sensor layout and measurement time.

56.3.2 Estimation of Experimental Modal Parameters The most exciting stage of the OMA can be stated as the stage of modal parameters estimation. Some commercial software is available for this calculation; the author prefers ARTeMIS Modal Pro for data analysis and modal parameter estimation [45]. This phase of the study mainly consists of five steps. • An analysis model is created; this is not the finite element model of the historic masonry structure. • Through accelerations, vibration measurements are uploaded to the software and assigned to the related degrees of freedom. • The analysis is carried out to obtain spectral density functions. • Estimation of the mode shapes and modal frequencies over spectral density functions can be done by using frequency domain decomposition (FDD), enhanced frequency domain decomposition (EFDD) analysis techniques, etc. Of the mentioned ones, EFDD analysis provides both modal frequencies and damping values, whereas only modal frequencies can be estimated by FDD analysis. • Validation of the mod shapes can be evaluated through modal assurance criteria.

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1689

56.3.3 Validation of Mod Shapes Modal assurance criteria (MAC), a statistical indicator, is used for validation. MAC is generally used to match the forms derived from experimental models and analytical models; it is also used to examine the mode shapes’ consistency. However, it can also be used to assess the consistency of the calculated mode shapes for the same system from different data sets. It is easy to implement and does not require the estimation of system matrices. MAC takes numerical values from 0 to 1. While the value expresses the shape of the entirely consistent mode, values close to 0 indicate that the modes are discrete [46]. If the two different mode shape vectors are compatible, the modal assurance matrix’s diagonal terms must be 1, and the other terms should be 0. The fact that some of the diagonal terms receive values that are considerably lower than the value of 1 indicates that the intersecting modes are very different from each other, and the resulting mode shape suggests the need to reconsider the shape vectors [47].

56.4

Possible Applications of OMA for Historic Masonry Structures

OMA can experimentally estimate the dynamic behavior parameters of a historic masonry structure. It is purely nondestructive and provides no damage to historic masonry. OMA results of historical masonry structures can be used for different purposes, some of which are briefly stated in the following part of the text.

56.4.1 Structural Model Calibration Theoretical modal analysis of a historic masonry building makes the dynamic identification of the system. It provides information about mod shapes, modal frequencies, and damping values of the structural model. On the other hand, OMA does the same; but for the real structure. Thus, the researcher may have information about dynamic behavior characteristics for both models and the corresponding real structure to make the comparison. This a quite valuable technical information that can be used to upgrade the structural model. Model calibration/upgrading to calculate the experimental modal parameters theoretically is becoming a common method to obtain a more reliable structural model for further structural analysis [48, 49, 50]. There are three major applications for theoretical model updating to reach experimental mode shapes and modal frequencies by this content. To reassign the mechanical properties of the masonry. Meaning that masonry mechanical properties are changed, and modal analysis is carried out until calculating the experimental modal parameters as closely as possible:

1690

D. Okuyucu

• To check the geometry and dimensions and make the changes, if necessary • To restudy and update the support conditions of the model If theoretical and experimental mod shapes well correspond, but related frequencies differ, reassigning the masonry’s mechanical properties is generally preferred for model calibration. Yet, this is a trial and error process. It needs the effort and sometimes engineering intuition of the analyzer. This implies the need for studies on the definition of mechanical properties of historic masonry. Some promising studies exist; but seem to be not enough [30, 51, 52]. If theoretical and experimental mod shapes differ, then it is suggested to check the geometry, dimensions, and assigned support conditions. The difference in mod shapes may also be an indicator of structural damage. In this way, OMA can also be preferred as a method for damage detection of historic masonry structures.

56.4.2 Structural Health Monitoring and Damage Detection Pulse is the heart rate or the number of times the heart beats in one minute. Heart rates vary from person to person. If the characteristic pulse of a person changes, it may be an indicator of a health problem. This is analogous to the structures. Mod shapes and modal frequencies are specific to a building, so the estimation of these parameters is called dynamic identification. In OMA, sensor locations are pre-defined to take the structural response like medical doctors do to take the pulse. Acceleration records are analyzed, and modal behavior parameters of the structure are estimated like medical people do to take and to evaluate the pulse. Thus, can regularly be estimated OMA results be used for structural health evaluation of a building? Can changes in mode shapes and modal frequencies give technical information about damage? Why not? Due to ease in application and nondestructive manner, OMA is preferred for modal parameters estimation of historic masonry buildings. The most important advantage of the method is to use ambient vibrations to obtain dynamic characteristics. Forced vibrations may damage the historical masonry and generally not possible to excite the building by a shaker. Considering all, regularly performed OMA is becoming more popular for structural health monitoring of historic masonry [53]. The approach is straightforward. Changes in modal behavior parameters can help to diagnose a structural health problem, damage. It is even suggested as a controlling tool in the restoration process of heritage buildings. Any constructional touch may change the building’s mass and rigidity, and it can affect the modal parameters and overall structural behavior [54]. The method is widely used for structural assessment, evaluation of safety conditions, and damage detection of historic buildings [55–58]. The following part of the text provides some OMA applications from the author’s work.

56.4.3 Some OMA Application Examples from Author’s Research Over 90% of Turkey’s area spans on Anatolian peninsula. Anatolia is one of the oldest occupational sites of the earth and hosted a number of civilizations in history. These

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1691

civilizations left their fingerprints on this land employing historic structures and folkloric traditions. Besides, Anatolia is one of the most active seismic zones of earth and experienced several highly destructive earthquakes in the past. Conditionally and intentionally, historic masonry buildings are the main research area of interest for the author. Protecting and carrying this architectural heritage to the future and committing them to our next generations are major responsibilities from a civil engineering perspective. The study on a historic masonry starts with the creation of a finite element model of the building by assigning initial mechanical properties of masonry, and theoretical modal parameters are calculated. In situ modal parameters of the structure are estimated by OMA and compared to that of theoretical ones for structural analysis model verification. In this phase, upgrading the mechanical properties of masonry and the creation of new approaches for mechanical property estimation of historic construction in a nondestructive manner become a primary concern. This is one of the main research topics of the author. In the last decade, restoration and preservation of architectural heritage became one of the primary concerns of the ruling government in Turkey. Under related investment programs, numbers of historic masonry constructions are technically studied. Investigating in situ modal behavior parameters of the historic structure before and after restoration is no longer a common application in Turkey, especially for historic bridges. The author also has some field experience in this type of work. The following part of the text provides information about some of the author’s OMA studies on historical constructions.

Studies Related to Finite Element Model Verification of Historic Masonry Buildings The nonhomogenous nature of masonry construction makes structural model creation challenging. The macro modeling approach’s adoption makes some modeling and analysis issues easier, but requires homogenization of the masonry and assigning related mechanical properties [11]. The author focuses on the investigation of the mechanical properties of historic masonry by nondestructive experimentation. The use of concrete Schmidt hammer for material characterization of historic masonry and application of OMA for an overall evaluation of structural modal behavior to compare that of theoretical calculations can be stated as some nondestructive experimental studies on historic masonry. OMA results of the historic building are used to verify the structural analysis model and check the masonry’s assigned mechanical properties derived by applying various methods. Modal analysis results of some historic buildings studied by the author are presented in Table 56.1. The structures shown in Table 56.1 were analyzed to evaluate seismic behavior, and structural analysis model calibrations were done for this purpose. Besides, the use of concrete Schmidt hammer to define historic masonry’s mechanical properties was also studied. Table 56.1 summarizes the theoretical and experimental modal frequencies of a total of four historic masonry constructions in Turkey. All structures were modeled by adopting a macro modeling approach. However, the material properties of composite masonry were initially defined and assigned by different approaches.

1692

D. Okuyucu

Table 56.1 A summary of theoretical modal analysis and OMA results on historic structures studied by the author Structure

Structure

Yakutiye Madrasah, Erzurum-Turkey, fourteenth century (~1310) [51] Mod Theoretical Experimental Difference numbera frequency (Hz) frequency (Hz) (%) 3 15.33 14.84 3.3 4 16.60 15.19 9.3 5 16.89 15.92 6.1 Lala Mustapha pasha mosque, Erzurum-Turkey, sixteenth century (~1562) [59] 3 8.44 9.42 10.4 4 8.50 10.55 19.4 5 9.30 10.85 14.3

Structure

Kadana mosque, Erzurum-Turkey, eighteenth century (~1751) [30] 3 4.81 4.48 7.4 4 5.06 5.86 13.7 5 6.09 8.67 29.8

Structure

Değirmenliköy church, Erzincan-Turkey, nineteenth century (~1860) [52] 1 6.38 3.56 79.2 2 7.71 5.18 48.8 3 8.78 6.89 27.4

Mosques are constructed with minarets. The first and second mode shapes were associated with the minarets’ local movements, which do not represent the whole structure’s modal behavior. Hence, the first and second mode shapes were not taken into consideration during the comparison.

a

Theoretical values are the results of the first models which were not calibrated. Hence, the differences between theoretical and experimental modal frequencies can be used to evaluate the definition method’s material properties’ success. It should be noted that theoretical and experimental mod shapes were almost the same individually, for all historic masonry buildings. For all applications, MAC values were considered for the validation of mode shapes. Among the structures stated in Table 56.1, Yakutiye Madrasah and Lala Mustapha Pasha Mosque were studied first. In the case of material characterization, the walls were assumed to be constructed in three vertical layers, and the term of three-leaf wall was used for this application, as depicted in Fig. 56.5. Schmidt hammer test results and calculation of crack density parameter were used to define mechanical properties of the inner and outer layers of the wall, whereas literature sources were taken into consideration to assign material properties of the rubble layer. The calculated mechanical properties of historic masonry were assigned to the model, and modal parameters were calculated [51, 59]. In these applications, the concrete Schmidt hammer and its original calibration sheet for concrete were used for compressive strength estimation of stone masonry units. The maximum difference between theoretical and experimental modal frequencies was obtained to be

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1693

Fig. 56.5 Three-leaf wall analogy [59]. (a) Analogy. (b) Common historic masonry wall practice in Turkey

9.3% and 19.4% for Yakutiye Madrasah and Lala Mustapha Pasha Mosque, respectively. The theoretical modal analysis results were promising for Yakutiye Madrasah, but the gap between theoretical and experimental modal analysis results widened for Lala Mustapha Pasha Mosque. The same procedure was then followed for Erzincan Değirmenliköy Church to define masonry’s mechanical properties [52]. The results provided very high values of error between theoretical and experimental modal frequencies, especially for the first mod shape at the level of 79.2%. It was the vertical movement of the vaulted roof made of stone. By making a detailed inspection, it was defined that the vault was mainly constructed by using a lightweight limestone (γ ¼ 13kN/m3). Besides, the porous inner structure of wall stones affected the Schmidt hammer measurements. Hence, roof and wall material’s physical properties were studied again, and updated material properties were assigned to the model [1]. The abovementioned three applications and related literature underline the need for producing new calibration equations for concrete Schmidt hammer to use for natural stones. Before studying the historic Kadana Mosque, comprehensive experimental research was carried out to create calibration equations for natural stones to be used for concrete Schmidt hammer. First, natural stones were taken from a historical masonry structure under restoration; physical and mechanical properties were defined. These values were linked to concrete Schmidt rebound values, and some equations were derived for basalt (Eq. 56.2) and andesite (Eq. 56.3) type of rocks, where fc is the compressive strength of the stone and Rv is the average of rebound values measured by horizontal application [30]. fc ¼ 0, 00012  ðRvÞ3:43

ð56:2Þ

fc ¼ 1, 01  ðRhÞ0, 97

ð56:3Þ

Equations 56.2 and 56.3 were used to estimate the compressive strength of stone masonry units of Erzurum Kadana Mosque. Estimated mechanical properties of

1694

D. Okuyucu

stone units by Schmidt hammer tests were combined with code equations to obtain mechanical properties of composite masonry [20, 21]. The difference between theoretical and experimental modal analysis results for the first two modal frequencies was acceptable for this building. At the same time, it was high for the third mod shape, 29.8%, which was the lateral-torsional movement of the building. The results show that it does not seem possible to generalize the derived equations and follow procedures for historic masonry unless a sufficient number of applications are carried out. By these real site applications, another way of using OMA is studied. The method can also be used for the verification of proposed masonry material characterization approaches.

An Application Example for Damage Detection of a Historic Masonry Bridge Details of a case study are presented in this part of the text as an example of a historic masonry bridge’s damage detection using the OMA technique. The name of the architectural heritage is Pehlivanlı Bridge. The single-span masonry arch bridge was constructed over the Dumlu River in Erzurum, Turkey. The construction date is reported to be the beginning of the nineteenth century [60]. The span and height of the arch are 12.50 meters and 15.17 meters, respectively. The bridge was restored in 2015. During and after restoration, the views of the bridge are presented in Fig. 56.6. To evaluate the modal behavior of the bridge and to study the structural effects of restoration, OMA applications were carried out before, during, and after repair and the results are presented in Table 56.2. Table 56.2 also provides the theoretical modal analysis results calculated on the finite element model created at the beginning of the study. The results presented in Table 56.2 were primarily used to see the effect of restoration work on the bridge’s structural behaviors. Primary reparation was done for the main arch and its keystone. Wide crack-type damages on supporting walls of the bridge were also repaired. As the results are evaluated, an increase in the bridge’s experimental modal frequencies after restoration is seen. This shows the increase in the stiffness, since the bridge’s mass was almost the same before and after the restoration. Besides, theoretical and all experimental mod shapes corresponded well. The study was conducted up to this point as a follow-up on the bridge’s modal behavior for its restoration period. Furthermore, as an addition to OMA

Fig. 56.6 During and after restoration views of Pehlivanlı Bridge in Erzurum, Turkey

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1695

Table 56.2 Theoretical modal analysis and OMA results of Pehlivanlı Bridge [60]

Mod number 1 2 3 4

Mod shape explanation Lateral translational movement Lateral torsional movement Vertical translational movement Vertical torsional movement

Frequency (Hz) OMA Application Before During Theoretical restoration restoration 14.97 5.28 5.44

After restoration 7.42

30.66

7.35

7.62

15.70

30.86

10.09

10.55

17.95

39.49

13.23

14.81

19.46

Table 56.3 OMA results of Pehlivanlı Bridge – almost 2 years after restoration Mod number 1 2 3 4

Mod shape explanation Local lateral translational movement of bridge body at the connection point of the bridge and supporting wall on the side of Kireçli Village Lateral torsional movement Vertical translational movement with participation of the connection point of the bridge and supporting wall on the side of Kireçli Village Vertical torsional movement with participation of the connection point of the bridge and supporting wall on the side of Kireçli Village

Frequency (Hz) 7.51 10.87 15.72 19.98

studies presented in Table 56.2, almost 2 years after the restoration work finished, another OMA application on Pehlivanlı Bridge was carried out. The results of the last OMA application of the bridge are shown in Table 56.3. OMA results of Pehlivanlı Bridge depicted in Table 56.3 presented some surprising results, through evident changes in mode shapes, which were all verified by MAC values. The modal behavior of the bridge was changed. Three of the four mod shapes were associated with the local movements of the bridge’s connection point and the supporting wall on the side of Kireçli Village. This could be an indicator of damage to the bridge at the mentioned connection point. It should be underlined that on the day of vibration measurement, a detailed visual inspection was done on the bridge as usual and no damage was observed. However, only a few weeks after the last OMA application, the pictures presented in Fig. 56.7 were taken. The pictures presented in Fig. 56.7 show the damage at the bridge’s connection point and supporting wall on the side of Kireçli Village. The damage became visible only a few weeks after the last OMA test of the bridge, which indicated possible damage at the village side’s connection point. Mod shapes estimated by the last OMA can be stated as the signs of damage. They started inside and were not visible yet on the day of ambient vibration measurements. The bridge was then evaluated, and improper infill application of the supporting wall on the village side was detected

1696

D. Okuyucu

Fig. 56.7 Views of damage on Pehlivanlı Bridge at the bridge’s connection point and supporting wall on the side of Kireçli Village

to cause the damage and change of the mode shapes. The results of this application show that damage detection of historic masonry can be another purpose of using the OMA technique.

56.5

Concluding Remarks

Masonry construction is as old as mankind and still finds a wide range of applications worldwide. Older masonry buildings, coming from early times of history, are the survivors of many destructive effects. The protection and preservation of historic masonry structures are multidisciplinary types of work. The structural engineer takes a big responsibility to make condition assessment and study architectural heritage’s structural behavior. This is a challenging engineering problem due to the nonhomogenous structure and numbers of unknown historic masonry construction. Yet, current approaches to estimate in situ mechanical properties of composite masonry for macro modeling purposes procedures seem to be far away from general applications. More and more research is needed in this area. The topic of the chapter is OMA applications for historic masonry constructions. OMA is nondestructive experimentation that provides information about in situ modal behavior parameters. With this fashion, OMA is an applicable technique for historic masonry studies from a structural engineering point of view. OMA by ambient vibration measurements can be preferred for: • • • • •

Structural model calibration Verification of masonry material identification procedures Structural health monitoring Condition assessment Damage detection

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1697

Purposes of historic masonry buildings. As the number of case studies on OMA use for historic masonry constructions increases, the gaps in the related literature will be closed, and the lack of knowledge for practical applications will be eliminated. More studies on this topic are expected.

References 1. Aslay SE, Okuyucu D (2018) Evaluation of 1939 Erzincan earthquake dynamic behaviour of Erzincan Değirmenliköy church over the abscissa damage. In: Proceedings 36th general assembly of the European seismological commission, Malta, pp 578–579 2. Gülkan P, Yücemen MS, Başöz N, Koçyiğit A, Doyuran V (1993) Turkish seismic zoning map prepared according to newest data, technical report: 93:01. Earthquake Engineering Research Center of Department of Civil Engineering, Middle East Technical University, Ankara. (in Turkish) 3. Ambraseys NN, Finke CF (2006) The Seismicity of Turkey and adjacent areas (Trans: Koçak MU). TÜBİTAK Publications, Ankara 4. Haçin I (2014) 1939 great Erzincan earthquake, Cumhuriyet University. J Soc Sci 88:38–69. (in Turkish) 5. Aslay SE, Okuyucu D (2020) Technical evaluation of abscissa damage of Erzincan Değirmenliköy church. J Fac Eng Archit Gazi Univ 35(1):387–402. (in Turkish) 6. Gürüz K (2016) Madrasah vs. University. Ka Kitap, Istanbul. (in Turkish) 7. Erzurum Archive. http://erzurumarsivi.com/tarihi-eserler/cifte-minareli-medrese/. Open Access 2019 8. Tozlu S (2000) Earthquakes in the History of Erzurum, Proc. on Seminar on Natural Disasters and Earthquakes in Anatolia Throughout History, Istanbul. Istanbul University Faculty of Literature Research Centre of History, pp 93–119. (in Turkish) 9. Musgrove J, Fletcher B (1987) Sir banister Fletcher’s: a history of architecture. Butterworths, London 10. Davey N (1961) A history of building materials. Phoenix House, London 11. Lourenco PB (2002) Computations on historic masonry structures. Prog Struct Eng Mater 4(3): 301–319 12. Okuyucu D, Kazaz I, Çodur MY, Kocaman I, Savaş GK, Kara T (2018) An investigation on effects of traffıc induced ground vibrations on historic structures examples of Erzurum Lala Mustafa Pasha Mosque and Yakutiye Madrasah. In: Proceedings 36th general assembly of the European seismological commission, Malta, pp 580–581 13. Okuyucu D, Kazaz I, Çodur MY, Kocaman I, Savaş GK, Kara T (2018) An evaluation on effects of the high level of sound on historic structures examples of Erzurum Lala Mustafa Pasha Mosque And Yakutiye Madrasah. In: Proceedings 36th general assembly of the European seismological commission, Malta, pp 579–580 14. Kocaman I, Okuyucu D, Kazaz I (2019) Determination of historical building material properties with dynamic parameters: the case of Lala Paşa Mosque. Teknik Dergi 30(3):9125–9146. (in Turkish) 15. Kocaman I, Kazaz I, Okuyucu D (2018) Investigation of the structural behaviour of historical Erzurum Yakutiye Madrasah, Dokuz Eylül University Engineering Faculty. J Sci Eng 20(58): 36–51. (in Turkish) 16. Lourenco PB (1996) Computational strategies for masonry structures, PhD Thesis. Delft University of Technology 17. Rots JG (1991) Numerical simulation of cracking in structural masonry. Heron 36(2):49–63 18. Anthoıne A (1992) In-plane behaviour of masonry: a literature review. Report EUR13840 EN. Commission of the European Communities, JRC – Institute for Safety Technology, Ispra

1698

D. Okuyucu

19. Lourenço PB, Rots JG, Blaauwendraad J (1994) Implementation of an interface cap model for the analysis of masonry structures. In: Mang H et al (eds) Computational modelling of concrete structures. Pineridge Press, Swansea, pp 123–134 20. Turkish Republic the Ministry of Public Works and Settlement: Turkish Earthquake Design Code-TEC 2018, Ankara, 2018 (in Turkish) 21. European Committee for Standardization: Eurocode 6: design of masonry structures, Part 1.1: general rules for reinforced and unreinforced masonry structures, Brussels, ENV 1996-1-1, 1995 22. expin: Advanced Structural Control Company: http://www.expin.it/servizi/indagini-strutturali/. Open access 2019 23. Flores-Colen I (2009) Methodology to evaluate the performance in service of rendered façades from the perspective of predictive maintenance, PhD thesis. Instituto Superior Técnico, Lisbon. (in Portuguese) 24. Miranda LF, Rio J, Guedes JM, Costa A (2012) Sonic impact method – a new technique for characterization of stone masonry walls. Constr Build Mater 36:27–35 25. Fort R, Buergo MA, Perez-Monserrat EM (2013) Non-destructive testing for the assessment of granite decay in heritage structures compared to quarry stone. Int J Rock Mech Min Sci 61: 296–305 26. Galvão J, Duarte R, Flores-Colen I, de Brito J, Hawreen A (2020) Non-destructive mechanical and physical in-situ testing of rendered walls under natural exposure. Constr Build Mater 230: 116838 27. Roknuzzaman M (2017) B Hossain, I. Mostazid, R. Haque: application of rebound hammer method for estimating compressive strength of bricks, journal of civil engineering. Research 7(3):99–104 28. Brozovsky J (2012) Implementation of non-destructive impact hammer testing methods in determination of brick strength. Appl Mech Mater 174–177:280–285 29. McCann M, Forde MC (2011) Review of NDT methods in the assessment of concrete and masonry structures. NDT E Int 34(2):71–84 30. Özdoğan DB (2018) Investigation of the use of Schmidt hammer method for definition of material properties of historical masonry structures: application of Erzurum Kadana Mosque, MSc Thesis. Erzurum Technical University. (in Turkish) 31. Mendes P, Baptista MA, Lagomarsino S, Costa JP (2005) Structural and dynamic analysis of N. Sra. Do Carmo Church, Lagos Portugal. In: Proceedings EURODYN Sixth European conference on structural dynamics, Paris 32. Betti M, Vignoli A (2008) Modelling and analysis of a Romanesque church under earthquake loading: assessment of seismic resistance. Eng Struct 30:352–367 33. Casolo S, Sanjust CA (2009) Seismic analysis and strengthening design of a masonry monument by a rigid body spring model: the “Maniace Castle” of Syracuse. Eng Struct 31:1447–1459 34. Brincker R, Zhang L, Andersen P (2000) Modal Identification from the ambient response using frequency domain decomposition. In: Proceedings of the 18th international modal analysis conference, San Antonio-Texas 35. Schwarz B, Richardson MH (2001) Modal parameter estimation from ambient response data. In: Proceedings IMAC-XIX conference, Florida 36. Schwarz B, Richardson MH (2001) Post-processing ambient and forced response bridge data to obtain modal parameters. In: Proceedings IMAC-XIX conference, Florida 37. Herlufsen H, Møller N (2002) OMA of a wind turbine wing using acoustical excitation. Brüel & Kjær Application Note 38. Møller N, Brincker R, Herlufsen H, Andersen P (2001) Modal testing of mechanical structures subject to operational forces. In: Proceedings IMAC-XIX conference, Florida 39. Brincker R, Andersen P, Møller N (2000) Output-only modal testing of a car body subject to engine excitation. In: Proceedings 18th international modal analysis conference, San AntonioTexas 40. Ceroni F, Sica S, Pecce MR, Garofano A (2014) Evaluation of the natural vibration frequencies of a historical masonry building accounting for SSI. Soil Dyn Earthq Eng 64:95–101

56

Operational Modal Analysis Method for Historic Masonry Structures:. . .

1699

41. Compan V, Pachón P, Cámara M (2017) Ambient vibration testing and dynamic identification of a historical building: basilica of the fourteen holy helpers (Germany). Proc Eng 199:3392–3397 42. Pepi C, Gioffrè M, Comanducci G, Cavalagli N, Bonaca A, Ubertini F (2017) Dynamic characterization of a severely damaged historic masonry bridge. Proc Eng 199:3398–3403 43. Sancibrian R, Lombillo I, Sarabia EG, Boffill Y, Wong HL, Villegas L (2017) Dynamic identification and condition assessment of an old masonry chimney by using modal testing. Proc Eng 199:3410–3415 44. Jacobsen NJ, Andersen P (2008) OMA on structures with rotating parts. In: Proceedings ISMA noise and vibration engineering conference, Leuven 45. Artemis Modal Pro (2017) Operational modal analysis software. Erzurum Technical University 46. Pastor M, Binda M, Harcarik T (2012) Modal assurance criterion. MMaMS 2012-Proc Eng 48: 543–548 47. Yeniceli SC (2014) Ground vibration testing process and importance in aircraft development projects. In: Proceedings V. National Aviation and Aerospace conference, Kayseri. (in Turkish) 48. Foti D, Diaferio M, Gianoccaro NI, Mongelli M (2012) Ambient vibration testing, dynamic identification and model updating of a historic tower. NDT&E Int 47:88–95 49. Sánchez-Aparicio LJ, Riveiro B, González-Aguilera D, Ramos LF (2014) The combination of geomatic approaches and OMA to improve the calibration of finite element models: a case of study in saint Torcato church (Guimarães, Portugal). Constr Build Mater 70:118–129 50. Genç AF, Ergün M, Günaydın M, Altunişik AC, Ateş Ş, Okur FY, Mosallam AS (2019) Dynamic analyses of experimentally-updated FE model of historical masonry clock towers using site-specific seismic characteristics and scaling parameters according to the 2018 Turkey building earthquake code. Eng Fail Anal 105:402–426 51. Kocaman I (2017) Determination of the necessary material properties for the calculation of dynamic behaviour of historic masonry structures, MSc Thesis. Erzurum Technical University. (in Turkish) 52. Aslay SE (2018) Evaluation of 1939 Erzincan earthquake dynamic behaviour of Erzincan Degirmenlikoy church over the abscissa damage, MSc Thesis. Erzurum Technical University. (in Turkish) 53. Ramos LF, Marques L, Lourenço PB, De Roeck G, Campos-Costa A, Roque J (2010) Monitoring historical masonry structures with OMA: two case studies. Mech Syst Signal Process 24(5):1291–1305 54. Masciotta MG, Ramos LF, Lourenço PB (2017) The importance of structural monitoring as a diagnosis and control tool in the restoration process of heritage structures: a case study in Portugal. J Cult Herit 27:36–47 55. Gentile C, Saisi A (2011) Ambient vibration testing and condition assessment of the Paderno iron arch bridge (1889). Constr Build Mater 25(9):3709–3720 56. Conde B, Ramos LF, Oliveira DV, Riveiro B, Solla M (2017) Structural assessment of masonry arch bridges by the combination of non-destructive testing techniques and three-dimensional numerical modelling application to Vilanova bridge. Eng Struct 148:621–638 57. Bautista-De Castro Á, Sánchez-Aparicio LJ, Ramos LF, Sena-Cruz González J, Aguilera D (2018) Integrating geomatic approaches, OMA, advanced numerical and updating methods to evaluate the current safety conditions of the historical Bôco bridge. Constr Build Mater 158:961–984 58. Gentile C, Saisi A (2007) Ambient vibration testing of historic masonry towers for structural identification and damage assessment. Constr Build Mater 21(6):1311–1321 59. Kocaman I, Okuyucu D, Kazaz I (2017) Determination of the related material properties for calculation of dynamıc behaviour of historic masonry structures. In: Proceedings 6th international symposium on conservation and consolidation of historical structures, Trabzon. (in Turkish) 60. Okuyucu D, Kazaz I, Çodur MY (2016) Determination of historical bridge seismic behaviour parameters using operational modal analysis method. Scientific research project: 2013/015. Project Coordination Office, Erzurum Technical University, Erzurum. (in Turkish)

On the Role of Historical Research in the Structural Condition Assessment of Heritage Structures

57

Adriana Marra, Carlo Rainieri, and Giovanni Fabbrocino

Contents 57.1 57.2 57.3 57.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review of National and International Regulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Role of Knowledge in the Structural Analysis of Historical Heritage . . . . . . Historical Research for Earthquake Engineering Applications . . . . . . . . . . . . . . . . . . . 57.4.1 Convitto Mario Pagano in Campobasso (South Italy) . . . . . . . . . . . . . . . . . . 57.4.2 Carthusian Monastery of Trisulti in Collepardo . . . . . . . . . . . . . . . . . . . . . . . . 57.4.3 Villa d’Este in Tivoli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.5 Final Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1702 1703 1706 1708 1709 1711 1716 1718 1721

Abstract

Natural events and the degradation of materials cause the loss of historical built heritage worldwide. An attempt to provide a comprehensive answer to the need for structural safety and conservation has been made during the last decades in several European countries because of major earthquakes. Innovative approaches and recommendations have been included in relevant technical codes, whose revision and validation are ongoing. Progress in technology and technical knowledge deployed new and detailed tools for the quantitative analysis of heritage A. Marra (*) ITC-CNR, Construction Technologies Institute, National Research Council, L’Aquila, Italy e-mail: [email protected] C. Rainieri (*) ITC-CNR, Construction Technologies Institute, National Research Council, Naples, Italy e-mail: [email protected] G. Fabbrocino (*) Department of Biosciences and Territory, StreGa Laboratory, University of Molise, Campobasso, Italy ITC-CNR, Construction Technologies Institute, National Research Council, L’Aquila, Italy e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_57

1701

1702

A. Marra et al.

structures, but this has led to the need of a continuous and conscious interaction between different professionals despite their cultural background. This chapter deals with the role of historical research in the technical process regarding the structural assessment and design of works to ensure the stability and durability of historical structures exposed to seismic and other natural hazards. Historical analysis is crucial in the assessment of built heritage, and it makes it possible to draft accurate and effective in situ investigation plans, essential for the implementation of structural models and analyses. Some examples of historical analyses complying with the needs associated to structural performance assessment are illustrated to this aim.

57.1

Introduction

The growing demand for the safety of historical built heritage throughout Europe yielded the enhancement of traditional approaches and methodologies for conservation by using novel Information and Communication Technology (ICT) tools [1, 2]. This is the case of several Mediterranean countries which are located in high seismicity areas and host a large historical built heritage to be preserved. In those countries, seismic events stimulated the development and use of seismic protection technologies in historical structures. Thus, local construction techniques were adapted to integrate anti-seismic solutions whose efficacy was tested by new events [3]. However, the modern paradigms adopted in the field of structural engineering, the development of new materials and strengthening techniques, as well as the lack of memory of past events weakened the attention to the building history resulting in the extensive execution of invasive interventions on built heritage. Early structural codes of practice disregarded the peculiarities of existing structures and above all historical ones [4], and the design rules for new buildings were applied also to the performance assessment of built heritage. The primary concern of such an approach is the requirement of high performance levels, which are often incompatible with the structural characteristics of existing buildings. As a result, extensive and irreversible structural interventions were carried out, and they significantly affected the historical value and the safeguard of the artifacts. The lack of a specific approach to the design of structural interventions considering the inherent value of history, materials, and construction techniques of historical structures caused severe cultural losses [5]. Similar circumstances occurred whenever new techniques and materials appeared on the construction market; as an example, the extensive use of reinforced concrete has yielded significant damage to historical buildings due to inadequate durability and corrosion of steel rebars. The most recent Italian standards and regulations outlined a coherent path for the design of structural interventions considering the peculiar characteristics of historical structures. The recommended performance-based approach [6], in compliance with the Eurocodes [7], represents a more rational method for the design of structural interventions. However, it requires the concurrent action of very different skills, often distributed among several professionals, in order to achieve a given knowledge

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1703

level about the analyzed artifact and to identify the most suitable interventions. Accurate investigations concerning the historical evolution, geometry, degradation, and materials allow minimizing the uncertainties in the structural and seismic assessment of existing structures. Large uncertainties are indeed often associated to the complex geometry, the variability of material properties, the stratification of different construction techniques, and the hidden damage caused by past events. Therefore, the design and implementation of reliable conservative actions require an interdisciplinary approach. In order to ensure structural safety as well as to safeguard the historical built heritage, the following significant innovations have been introduced in the codes: (i) criteria to adapt the level of safety, which can be lower than for new buildings; (ii) physical models that are adequately representative of existing buildings; (iii) analytical and numerical models that quantify the response under static and dynamic actions; and (iv) identification of materials and construction techniques that can increase safety and preserve the values of historical buildings. The present chapter outlines the role of historical research in the analysis of the historical built heritage and its potential whenever an extensive knowledge of the artifact is required. The use of historical analysis as a tool for an effective planning of in situ investigations and for an accurate structural assessment preserving the historical and cultural value of the artifact is discussed, pointing out its role also for the setting of accurate structural models.

57.2

Review of National and International Regulations

The new awareness acquired in the assessment of built heritage has highlighted the importance of a multidisciplinary approach to conservation and safety issues. Therefore, the national and international regulatory framework has been strongly renewed. In particular, in Italy, the new Technical Standards for Construction [6] and the related so-called Circular 7/2019 [8] have been recently approved. Both documents report important changes and explanations on the evaluation of the structural performance of existing buildings, underlining the importance assumed by knowledge in the analysis. Three Knowledge Levels (KL), with increasing depth, are defined in order to overcome the uncertainties related to the analysis of existing buildings. The KLs directly derive from the study of the artifact and are related to history, geometry and defects in construction, knowledge of materials, and construction techniques. The KLs are also influenced by the state of conservation, that is to say, the effects of degradation and lack of maintenance, and by stratification of interventions and damage repairs over time. While this information can be obtained through laboratory and in situ investigations, including nondestructive as well as weakly destructive testing, after the Circular [8] an exhaustive knowledge of the artifacts through a different kind of analysis is required before executing experimental tests. These must be appropriately planned to fill the knowledge gap resulting from the historical and morphological dimensional analysis; moreover, they must investigate critical details for the type of analysis and considered loads. Thus, historical research plays a primary role in

1704

A. Marra et al.

identifying the evolution of the structure over time and the construction materials and techniques, and this is critical for the appropriate definition of the test plans aimed at the assessment of mechanical properties of materials and structural details minimizing the impact on the structure. The last update of the Italian seismic code NTC 2018 confirms this approach and provides additional details for the performance assessment of existing buildings. The “Guidelines for the evaluation and mitigation of seismic risk to cultural heritage” [9], drawn up in 2011 by the Ministry of Cultural Heritage, agree with the previously described framework, so they can be still considered valid for the analysis of the Italian listed heritage. These recommendations define a rational path to the selection of appropriate structural interventions. In particular, they trace a knowledge path to quantify partial confidence factors (CF) to modify the capacity parameters used in the analyses. Partial CFs are related to the KL reached after the extensive analysis of different building features and sources of information, such as geometric survey, identification of historical and construction peculiarities of the artifact, mechanical properties of materials, type of soil, and foundation. After a final judgment on the safety and conservation of the existing heritage structure has been reached through the knowledge path, specific guidelines to set representative models of the analyzed historical structure can be found in the Recommendations. Those recognize that design and construction criteria, based on the rules of art and on direct observation of the behavior of similar structures, are different from those used for modern buildings. As a result, the behavior of historical buildings is not easily replicable because of the unique features of each artifact. Three levels of assessment (AL1, AL2, AL3), associated to increasing knowledge about structural details, are defined for seismic performance evaluation, ranging from the analysis of individual structural units or single elements – macro elements – susceptible to local mechanisms up to global methods of structural modeling and analysis [11]. At the international level, the ISO 13822 standard [10] deals with the evaluation of existing buildings. A hierarchy scheme is defined (Fig. 57.1) depending on the final goal of the assessment (safety or safeguard). Different analysis steps, associated to increasing levels of detail, are defined. Visual inspections and analysis of design documents are the fundamental steps of a preliminary assessment aimed at identifying structural weaknesses. In this phase the collected data must be verified or updated to include information about past interventions affecting the structure. Particular attention must be focused on the occurrence of significant environmental actions, including seismic events, changes in soil conditions, and misuse of the structure. After the preliminary data collection on the artifact, relevant uncertainties affecting the evaluation of the structural performance can be identified and an appropriate inspection plan can be set to enhance the understanding of the structure and the analysis models. The ISO 13822 standard remarks how historical analysis can provide a large amount of information about existing buildings. Moreover, appropriate partial safety factors, or additional random variables, must be considered in structural analysis to account for the modeling uncertainties, in agreement with the Italian codes and guidelines.

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1705

Fig. 57.1 Flowchart for the general assessment of existing buildings. (Adapted from [10])

The ISO 13822 also includes an annex focused on heritage structures reporting additional prescriptions and recommendations for the application of the standard to buildings with cultural and historical value. When heritage structures are considered, the assessment takes account of the structural performance as well as the value of the structure as a cultural resource. The assessment phase is similar to ordinary structures (Fig. 57.1). However, the role of historical analysis and a multidisciplinary team of professionals (engineers, architects, historians, archaeologists, conservators, etc.) is even more important. In particular, the historical analysis aims at identifying the nature of the original construction, all subsequent alterations, and any significant event that caused structural damage. Historians or archaeologists often report this information in a general document. The role of the engineers is to cooperate with and support those professionals in the identification and interpretation of documents relevant to the structural assessment and conservation of heritage buildings. As with ordinary buildings, the structural analysis should consider the uncertainties resulting from incomplete data, damage and hidden defects, the limited knowledge of the ancient structural system, and the variability of material properties. Thus, it might be necessary to consider multiple structural models and execute additional (destructive) tests if the models are not sufficiently reliable.

1706

A. Marra et al.

The recommendations of the International Council on Monuments and Sites [12] are another reference document for the identification of repair methods suitable for architectural heritage. These recommendations also define an analysis methodology similar to healthcare procedures: anamnesis, diagnosis, therapy, and control [13]. The approach is once again multidisciplinary, thus confirming the need for different complementary skills to approach the analysis and investigation of heritage structures and to select appropriate interventions fulfilling the objective of safeguarding cultural heritage.

57.3

The Role of Knowledge in the Structural Analysis of Historical Heritage

The knowledge about historical heritage has an important role in the evaluation of structural performance (anamnesis and diagnosis) and in the identification of the strategies to implement for conservation and maintenance (therapy and control). It aims at filling the gaps regarding the construction and modifications of the artifact, reducing the uncertainties about geometry, material properties, construction techniques, and boundary conditions. The knowledge path outlined by the Italian and international regulations supports and guides the technicians towards the selection of the most appropriate analysis methods based on the collected data and information. However, the knowledge path is too often considered as a simple data acquisition phase, ending when the most significant data and information have been collected. This might be a limitation, because it does not exploit the multidisciplinary and dynamic process behind the knowledge of historical heritage. It is multidisciplinary because it involves different skills and disciplines, and it is dynamic because it combines the different information collected to properly select the performance evaluation method and identify the interventions able to ensure the safety and safeguarding of the artifact. The knowledge path is based on the scientific-systematic method. It is characterized by an ordered sequence of technical and scientific actions and activities to achieve conservation and safety (Fig. 57.2). The knowledge path is usually organized in the following five phases: 1. Identification of the construction in its territorial context and collection of information about the construction and urban history of the area where it is located 2. Review of the historical and constructive process 3. Description of the geometry of the building, including phenomena of alterations or disruptions due to past or ongoing events 4. Recognition of materials, state of degradation, and mechanical properties 5. Identification of the load-bearing elements of the structure. The historical and constructive analysis of the building heritage is definitely a critical moment in the knowledge path. It leads to the identification of the specific historical features of the building in terms of transformations that occurred over time, materials, and construction techniques.

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1707

Fig. 57.2 The path of knowledge. (Photo by authors)

The historical analysis requires the preliminary collection of all published documents concerning the building as well as the urban or territorial area where it is located. Examining archival documents and bibliographic and iconographic sources allows to retrace the history of the building from its construction to its current state. A detailed analysis of all the sources of information highlights the transformations, restorations, or partial destructions that affected the historical artifact after events of either modest or significant importance. Even dates concerning minor events and facts become significant in the historical analysis, because thanks to them it is possible to draw up a detailed chronological report of the inspected sources. In addition to the historical knowledge derived from written sources, specific analysis methodologies usually applied in archaeology can support the analysis of the building material, which becomes a document and a direct source to retrace the history of the built heritage. Examining materials and construction techniques, the stratification of interventions over time can be identified [14]. At the same time, the availability of these data fills the gaps between undocumented events, the evidence of which can be recovered only through a reciprocal study of several sources. The scientific and technical progresses in the field of archaeology and, in particular, in the identification of materials and construction techniques used in a given historical period and in a specific territorial area [15, 16] make possible the use of material

1708

A. Marra et al.

sources to trace the history of an artifact. In addition, the information that can be found in the restoration codes of practice, drawn up in the Italian context for different historical centers, allow for the codification of different building features and provide guidelines for the interventions [17–19]. The selection of appropriate interventions is based on an extensive knowledge and a systematic comparison between the construction rules of ancient buildings and the new construction technologies. The analysis of historical data yields the constructive features of the buildings, but further investigations must be carried out to validate the obtained results. Several diagnostic investigation techniques, such as infrared thermography, sonic tests, and endoscopy, can be used to support and validate the historical analysis. These techniques allow to identify the types of masonry and their attribution to a specific phase of construction. Thermography, for example, can detect structural issues and support the evaluation of the condition of masonry structures and the identification of the types of horizontal structures, without affecting their state of preservation [20, 21]. Nondestructive testing techniques provide information useful for the qualitative evaluation of the mechanical properties of materials. Similar information can be obtained from technical documents, such as reports, specifications, and contracts referring to recent interventions. The collected information, validated by the diagnostic tests, can be compared with reference literature values and code provisions. In the knowledge path, the comparing of the information deriving from the historical analysis of built heritage with that from the dimensional analysis based on direct or indirect survey is crucial. Indeed, the identification of buildings and structural units starts from the graphical representation of historical transformations in plan or elevation, showing the artifact as a whole and its component elements. Moreover, this comparison makes possible the discrimination between the original structures and those added over time and the identification of the construction peculiarities and structural characteristics of load-bearing elements. These results allow the preliminary diagnosis of the current state of the artifact and guide the research towards further analyses aimed at completing the knowledge framework. The in situ or laboratory investigations might often reveal unknown features about the history of analyzed artifacts [22]. The present discussion remarks how the analysis and accurate representation of data coming from history, survey, and material sources is a fundamental step for critical judgments about the cultural, historical, and artistic value of the historical structure in its current form and in its component elements. At the same time, the structural characteristics and stability conditions of the structure can be defined through data integration, and a structural model adapted to the construction type and to the performed analysis can be set [11].

57.4

Historical Research for Earthquake Engineering Applications

The knowledge of built heritage is achieved through different phases that are strongly related to each other even if they are characterized by independent and autonomous properties. These stages, previously described and which can be further

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1709

summarized in historical, constructive, and dimensional analysis, are mutually interconnected and reciprocally interacting. The purpose of historical analysis, as previously mentioned, is the knowledge of the events that gave rise to the built heritage, that is, the synthesis and the material expression of a complex system in which history and technological culture merge. This research is mainly directed towards archives and aims to collect all the material that has been published on the building, going back to the moment of its construction, in order to draw up a chronological report and an iconographic synthesis. This information is correlated to the result of the dimensional analysis, and, therefore, the transformation phases of the artifacts are represented in plans and elevations. The result of this operation is further enhanced when the information deriving from the constructive and material analysis is introduced in the process, providing valuable data to conclude the path for the safety and safeguard of the built heritage. In the following, some case studies are presented to emphasize the role that history plays in the knowledge process, necessary for the assessments of the structural performance of built heritage, with an attempt to remark how it supports the achievement of an exhaustive level of knowledge. They refer to representative listed buildings of the Italian heritage; however, the adopted approach appears to be appropriate for any existing structure, including the minor ones (see, for instance, [14]).

57.4.1 Convitto Mario Pagano in Campobasso (South Italy) Following the L’Aquila earthquake in 2009 and the collapse suffered by the Convitto (National Boarding School) in the same city, attention was paid to the structural performance of the corresponding Institution in Campobasso. A seismic vulnerability assessment was required for the Convitto Mario Pagano (Fig. 57.3a), located in the administrative center of the Molise region in Southern Italy (Fig. 57.3b). The building belongs to a network of new schools built all over the Country after 1860, the year of the Italian unification; the typological peculiarities of these constructions and a comparative analysis between them have been the subject of a comprehensive publication [23]. As far as the Convitto in Campobasso is concerned, a specific research at the National Archive of Campobasso and at the Historical Archive of the Convitto Mario Pagano was carried out during the evaluation of the seismic vulnerability of the complex. It is an interesting example of the role of historical research that supports structural diagnostics and analysis. Several structural deficiencies due to relative displacements between walls and evidence of soil deformations were in fact documented and clearly reported in the technical documents found during the research (Table 57.1). This is the case of some crack patterns found in the area of the building reported in Fig. 57.3c, whose recurrence is confirmed by historical analysis and is quite clearly associated with the weak performance of foundations. Indeed, the 1887 report highlights the presence of the crack patterns due to defects in the building construction; later on, the 1893 report connects the cracks to the changes of the structure made during the development of the surrounding urban area. Furthermore, historical documents confirm

1710

A. Marra et al.

Fig. 57.3 Convitto Mario Pagano in Campobasso: (a) building location; (b) location of Molise Region; (c) general view of the school during the works carried out at the beginning of the twentieth century (Historical Archive of the Convitto Mario Pagano), (d) section on the staircase (National Archive of Campobasso, Genio Civile I, b. 16, f. 11); (e) project table for the ceiling of the Aula Magna (National Archive of Campobasso, Genio Civile I, b.18, f. 23)

that the profile of the soil around the constructions suffered several changes and therefore had a negative influence on the original structures of the building. Consequently, interventions to repair the damage observed have been made also in the twentieth century (Table 57.1). It is also worth noting that the quality and accuracy of many drawings made it possible to perform a detailed structural survey of the building and enabled a high level of knowledge of structural details, like those reported in Fig. 57.3d and e, referring to the decorated floor of the refectory.

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1711

Table 57.1 Main historical events related to the Convitto Mario Pagano in Campobasso Year 1895–1899

1888–1903 1930–1931 1972–1978 1990s

Events The part of the building facing Via Veneto, previously Via Novelli, was completed with the construction of a paved embankment at the height of the current main entrance of the building Cracks related to foundation problems were found The initial configuration of the embankment was modified carrying out an excavation and, later, building the garden in front of via Mazzini Construction of the reinforced concrete roof space. Restoration of the wall along Via Veneto and of the semi-circular loggia located near the main access to the Convitto, which was affected by overturning towards Via Mazzini

57.4.2 Carthusian Monastery of Trisulti in Collepardo In this section, some aspects of a wide research project focused on historical buildings hosting Italian Museums are reported. Attention is paid to the Carthusian Monastery of Trisulti in Collepardo, in the Lazio region. It is a different kind of historical construction compared with the previous one, since it is by far older and more important from the historical and cultural point of view. An extensive research, based on many bibliographic and iconographic sources, has been carried out to achieve a comprehensive knowledge about artistic, architectural, and technical history. Figure 57.4 provides an overview of the complex from a geographical and architectural point of view, and it shows some of the documents and pictures found in national and specialized libraries. The documents from the archives and library of the Carthusian monastery, concerning the history and transformations of the artifact as well as the artworks located in the different buildings and the artists who worked there, were enriched with the technical reports found in the national archive of Frosinone. The latter documents mainly concern the reports and the bill of quantities drawn up for the restoration and consolidation works carried out after the 1915 Avezzano earthquake (Fig. 57.5), as well as the works planned by the Soprintendenza between 1972 and 2009 [24, 25]. Table 57.2 shows the detailed chronology of the events that involved the monastic complex over the centuries. Historical research continued with the analysis of material sources, namely, construction materials, such as the types of masonry, used in the different buildings that characterize the complex. The information was collected in forms and other graphic supports, i.e., plans and elevations; comparing in situ surveys with the historical archaeology documents [16] and with the information included in the technical reports found in the archives (Fig. 57.5d) provided useful information on the techniques and construction materials used on the territory in a specific time.

1712

A. Marra et al.

Fig. 57.4 The Carthusian monastery of San Bartolomeo in Trisulti: (a) location of Lazio Region; (b) Carthusian monastery location; (c) general view (photo by authors). Photos of some iconographic sources from Historical Archive of the Carthusian monastery of Trisulti: (d) Maison de l’Orde des Chartreux (1895); (e) project of the new refectory by Tommaso Catrani (1763); (f) the church façade in the project of Paolo Posi (1768). (Photos by authors)

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1713

Fig. 57.5 The Carthusian monastery of Trisulti: (a) picture of the pharmacy; (b) reinforced concrete members roof above the pharmacy; (c) detail of a node and (d) notes of the Impresa Levantesi Umberto di Quintino (National Archives of Frosinone). (Photos by authors)

1714

A. Marra et al.

Table 57.2 Main historical events related to the Carthusian Monastery of Trisulti Year 1204 1208 1634 1662

1740

1749 1768 1776–1770 1770 1936 1958 1979 1985

Events Pope Innocent III started the construction of the church and the monastery The Carthusian religious occupied the new monastery The visiting fathers ordered the Carthusian religious of Trisulti to enlarge the Sacristy A plan preserved in the archive, dated 22 May 1662, documents how the pharmacy was to be built (as it appears today) raised on a previous one on the ground floor. The project was modified and completed after 1763–1770 The prior Pietro Paolo Bedini started several works that transformed the Carthusian monastery: the layout of the great cloister; the layout of the building complex of the entrance; the raising of the pharmacy The new bell tower in the small cloister was finished Demolition of the church facade and reconstruction according to a design by Paolo Posi Renewal of the refectory based on a project by Tommaso Catrani Building of new structures and embellishments in the Palace of Innocent III Interventions to repair the damage caused by the 1915 earthquake Demolition of the eighteenth-century structures in the Palace of Innocent III Consolidation of the wall structures of the refectory Consolidation and restoration work to repair the damage caused by the 1984 earthquake

The systematic organization of construction and transformation data was combined with field inspections and analyses, supported by nondestructive diagnostic techniques. Thermographies and endoscopies, with a limited impact on the structures under investigation, allowed the identification of the constructive characteristics of structural elements [26]. In fact, the hidden features of the masonry were identified through the collection of information about homogeneity of the material (masonry) used during the construction and restoration works and the definition of qualitative parameters for the evaluation of the structural performance. The historical analysis provided, also in this case, an important support to the structural diagnosis and assessment. The several iconographic sources allowed to understand the destructions and partial extensions of some structural units of the complex and therefore to characterize the structural deficiencies of some elements. For instance, the southeastern area of the large cloister exhibits a widespread crack pattern that makes a significant part of the artifact susceptible to local collapse mechanisms. This crack pattern could have been caused by the new constructions to the building, as shown in Fig. 57.4c and d. At the same time, the works carried out on the pharmacy after the seismic events of 1915 and 1984 (Figs. 57.5 and 57.6) caused considerable alterations to structural and nonstructural components, especially in the Balbi room. Seepage of water into the roof spaces caused widespread degradation of the main beams and the false ceilings.

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1715

Fig. 57.6 The Carthusian monastery of Trisulti: (a) picture of the pharmacy; (b) results of passive thermography. (Photos by authors)

The adopted approach, exploiting the correlation between historical analysis results and in situ investigations, allowed to recognize the structural elements and original construction (Fig. 57.7). The comparison between written documents and material sources enhanced the knowledge about the construction and transformation phases of the complex and identified issues in need of further investigations. The architectural complex was divided into structural units for safety assessment. Nondestructive or weakly destructive investigations, indeed, allowed to resolve those uncertainties that could negatively affect the structural modeling [27] and the evaluation of the structural performance [11, 28].

1716

A. Marra et al.

Fig. 57.7 Example of reconstruction of the historical-constructive evolution, Carthusian monastery of Trisulti. (Photo by authors)

57.4.3 Villa d’Este in Tivoli In the framework of the research project mentioned in the previous section, the Villa d’Este complex in Tivoli, in the Lazio region, was also analyzed (Fig. 57.8). The Villa, which is an important example of the Italian Renaissance and is included in the UNESCO Heritage List since 2001, has suffered several changes over the centuries and especially following the bombing of the Second World War [29]. All the information on the history of the complex has been collected from different sources consulted in National archives and libraries, as well as from the technical reports found in the museum’s technical office (Table 57.3). The presence of large frescoed and decorated surfaces (Fig. 57.8e) has required the use of nondestructive investigation techniques as tools to support the historical analysis. Passive thermographies and endoscopies on partially damaged or unfinished structural elements were carried out to validate the information derived from written sources and the identification of the constructive characteristics of the structural elements (Fig. 57.9). At the same time, direct in situ investigations were also carried out. During these operations, the information was collected in forms and other graphic supports (i.e., plans

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1717

Fig. 57.8 Villa d’Este in Tivoli: (a) location of Lazio Region; (b) location of Villa; (c) view of Great Loggia; (d) view of a lateral façade; (e) inner view of the “Stanza della Gloria”. (Photos by authors)

and elevations) in order to obtain useful data to complete the knowledge framework achieved with the analysis of documentary sources and the diagnostic surveys. The historical analysis of Villa d’Este clarified the relationship between the new construction and the remains of the ancient Roman villa, located on the site before the new villa was built [30], or the alignment and the connection between the ancient walls and those rebuilt after the aerial attack during the Second World War. Thanks to the correlation between the different research results pursued and assessed as part of the historical analysis, it was possible to improve the knowledge about the construction (Fig. 57.10) and transformation phases (Fig. 57.11) of the complex. Moreover, useful data were obtained for the most appropriate structural

1718

A. Marra et al.

Table 57.3 Main historical events related to Villa d’Este Year 1256

1550

1555 1560–1563 1566–1568

1628–1641 1670–1672 1672–1686 1804–1810 1851–1896 1922–1923 1944

1946–1948

1957–1960

Events Pope Alexander IV abolished the Benedictine monastery (built on the site of a laterepublican Roman villa) annexed to the church of Santa Maria Maggiore (probably dating back to the fifth century), giving part of it to the Franciscans, who brought about considerable transformations Cardinal Ippolito II d’Este, appointed as Governor of Tivoli, rented and bought the land and houses of the medieval quarter located in the so-called “Gaudente Valley,” for the construction of the Este palace, which would incorporate both the town hall and the Franciscan monastery The first design by Pirro Ligorio for the palace can be attributed to this date Work was resumed on the villa and the painting decorations of the interiors were started The final layout of the palace can be dated back to 1566 with a more regular front and with angular towers: in 1566–1567 the central body was enlarged from the initial seven axes to nine, and in 1568 the construction of the two angular towers was started Under the direction of the architect Francesco Peperelli, some restoration and maintenance works were undertaken The architect Mattia de’ Rossi suggested some improvement and decoration projects of the villa Refurbishment works continued and some maintenance work was undertaken Some improvement works were realized on the windows and doors, on the roofs, and on some masonry walls Some restoration works on deteriorated parts of the building The villa was handed to the Italian ministry. Restoration works on the decorations of the building On 26 May, an aerial attack damaged the northeast side of the building. Some restoration work was then realized on the building, repairing the walls, rebuilding wooden and reinforced concrete floors, as well as the roofs Some restoration works on the masonry walls and some consolidation works inside the building, as well as some frescoes, damaged by the war. The lacunar ceilings in the five rooms of the north-east side of the building destroyed by the bombing were also reconstructed Some ordinary maintenance interventions were performed together with some consolidation interventions as well as paving and finishing works in the internal court

modeling and analysis and, last but not least, the identification of issues in need of further investigations [29].

57.5

Final Remarks

In the context of the seismic assessment and design of structural interventions on heritage structures, the knowledge obtained by comparing the results from different analyses points out issues that are not easy to detect. The correlation between data

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1719

Fig. 57.9 Villa d’Este in Tivoli: (a) graphic reconstruction of the facade on the gardens; (b) results of passive thermography. (Photo by authors)

Fig. 57.10 Historical-constructive evolution, Villa d’Este in Tivoli. (Photo by authors)

derived from several analyses and sources (historical, bibliographic, iconographic, dimensional analysis – complete with deformation and cracking – and materials and construction techniques) drives the research in other directions in order to find further elements to complete the knowledge base. Moreover, the data obtained during this first phase of the knowledge path must be related to results

1720

A. Marra et al.

Fig. 57.11 Restoration works, Villa d’Este in Tivoli. (Photo by authors)

deriving from diagnostic investigations. Among these, all the techniques that preserve the integrity of the object, because they do not require any contact with it, should be preferred. The correlation between different data provided a detailed and reliable characterization of the structural elements and useful information for the identification of the structural element aggregation, for the identification of single macro elements, and for the qualitative characterization of the mechanical parameters of the materials. As a result, it is possible to define models adequately representative of the structural behavior of the building to assess its performance and safety. It is therefore clear that the proper identification of the transformation phases of the structural system and the identification of the discontinuities and constructive inhomogeneities, related to the damage caused by natural and manmade events, provide the main features for the evaluation of the built heritage. In fact, they provide a reliable interpretative model of the construction pursuing the aim of conservation. Although the illustrated applications refer to assets located in seismic areas, the approach and role played by historical research in the structural assessment are effective in all territorial contexts and for any type of construction. These examples also confirm that multidisciplinary skills and advanced technological tools are crucial to achieve a reliable knowledge about the structure, which is essential to guide repair and strengthening works in compliance with the safety and preservation needs of the built heritage. Acknowledgments The authors are grateful to ARCUS spa and MIBACT for their research contribution provided for the analysis of the Carthusian monastery of Trisulti in Collepardo (Fr) and Villa d’Este in Tivoli (Rome) within the project “Seismic Assessment of National Museums.”

57

On the Role of Historical Research in the Structural Condition Assessment. . .

1721

References 1. Bruno S, Fatiguso F (2018) Building condition assessment of built heritage in Historic Build Information Modeling. Int J Sustain Dev Plan 13:36–48 2. Percy K, Ward S, Santana Quintero M, Morrison M (2015) Integrated digital technologies for the architectural rehabilitation & conservation of Beninn Bhreah Hall & Surrounding Site, Nova Scotia, Canada. ISPRS Ann Photogram Remote Sens Spatial Inf Sci 2(5) W3:235–241 3. Ferrigni F (2005) The local seismic culture. In: Ferrigni F, Helly B, Mauro A, Mendes Victori L, Pierotti P, Rideaud A, Teves Costa P (eds) Ancient buildings and earthquake. Reducing vulnerability of historical built-up environment by recovering the Local Seismic Culture: principles, methods, potentialities. EDIPUGLIA, Bari, pp 199–214 4. ICOMOS (1964) The Venice Charter. International charter for the conservation and restoration of monuments and sites. https://www.icomos.org/en/resources/charters-and-texts 5. Borri A, Corradi M (2019) Architectural heritage: a discussion on conservation and safety. Heritage 2(1):631–647 6. NTC (2018) Ministerial Decree January 17, 2018. Updating of Technical Standards for Construction. G.U. n. 42 20/02/2018, Roma, in Italian 7. EN 1998-1 (2003) Eurocode 8: design of structures for earthquake resistance – Part 1: general rules, seismic actions and rules for buildings. CEN TC 250, Brussel 8. Circular 7 (2019) Guidelines for the application of the updating of the technical standards for construction provided by the Ministerial Decree of 17 January 2018. G.U. n. 35 11/02/2019, Roma, in Italian 9. Recommendations PCM (2011) Directive of the Prime Minister, 09/02/2011. Guidelines for the assessment and the mitigation of seismic risk of cultural heritage with reference to Italian NTC2008. G.U. n. 24 29/01/2011, Roma. in Italian 10. ISO 13822 (2010) Bases for design of structures – assessment of existing structures. ISO TC98/ SC2, Geneva 11. Fabbrocino G, Brigante D (2018) Approccio metodologico all’analisi sismica della Certosa. In: Fabbrocino G, Savorra M (eds) La Certosa di Trisulti. Silvana Editoriale, Milano, pp 190–199 12. ICOMOS (2003) ICOMOS charter – principles for the analysis. Conservation and Structural Restoration of Architectural Heritage 13. Lourenço PB (2006) Recommendations for restoration of ancient buildings and the survival of a masonry chimney. Constr Build Mater 20:239–251 14. Parenti R (1992) Fonti materiali e lettura stratigrafica di un centro urbano: i risultati di una sperimentazione non tradizionale. Archeologia Medievale: cultura materiale, insediamenti, territorio XIX:7–62 15. Aveta A (1987) Materiali e tecniche tradizionali nel napoletano. Note per il restauro. Arte Tipografica, Napoli 16. Fiorani D (1996) Tecniche costruttive murarie medievali. Il Lazio meridionale. “L’Erma” di Bretschneider, Roma 17. Giuffrè A (ed) (1993) Sicurezza e conservazione dei centri storici: il caso Ortigia: codice di pratica per gli interventi antisismici nel centro storico. Laterza, Roma 18. Marconi P (1997) Manuale del recupero del centro storico di Palermo. Flaccovio Editore, Palermo 19. Caravaggio P, Meda A (2004) Manuale del recupero di Castel del Monte. DEI, Roma. in Italian. 20. Binda L, Lualdi M, Saisi A, Zanzi L., Gianinetto M, Roche G (2003) NDT applied to the diagnosis historic buildings: a case history. Proceedings of 10th international conference structural faults & repair. London 21. Casapulla C, Maione A, Argiento LU (2018) Infrared thermography for the characterization of painted vaults of historic masonry buildings. Int J Struct Glass Adv Material Res 2:46–54 22. Rainieri C, Marra A, Rainieri GM, Gargaro D, Pepe M, Fabbrocino G (2015) Integrated non-destructive assessment of relevant structural elements of an Italian heritage site: the Carthusian monastery of Trisulti. J Phys Conf Ser 628(1). https://doi.org/10.1088/1742-6596/ 628/1/012018

1722

A. Marra et al.

23. Savorra M (2011) Il convitto nazionale Mario Pagano di Campobasso. In: Mangone F, Tampieri MG (eds) Architettare l’Unità: architettura e istituzioni nelle città della nuova Italia, 1861–1911. Paparo edizioni, Napoli, pp 243–248 24. Savorra M (2018) La storia, la costruzione, i committenti. In: Fabbrocino G, Savorra M (eds) La Certosa di Trisulti. Silvana Editoriale, Milano, pp 31–45 25. Marra A (2018) I restauri del Novecento. In: Fabbrocino G, Savorra M (eds) La Certosa di Trisulti. Silvana Editoriale, Milano, pp 46–55 26. Rainieri C, Rainieri GM, Gargaro D (2018) Diagnostica non distruttiva. In: Fabbrocino G, Savorra M (eds) La Certosa di Trisulti. Silvana Editoriale, Milano, pp 184–189 27. Gerbino S, Celiento A (2018) Il disegno e la modellazione come strumento di conoscenza. In: Fabbrocino G, Savorra M (eds) La Certosa di Trisulti. Silvana Editoriale, Milano, pp 170–177 28. Rainieri C, Marra A, Fabbrocino G (2014) On the estimation of the fundamental modal properties of Italian historical masonry towers. Ingegneria Sismica Int J Earthquake Eng 31(3):1–15 29. Marra A, Brigante D, Rainieri C, Fabbrocino G (2016) Structural characterization and performance assessment of the Villa d’Este Palace in Tivoli. In: Modena C, da Porto F, Valluzzi MR (eds) Brick and Block Masonry – trends, innovations and challenges. CRC Press, Padova, pp 645–652 30. Fratini G, Moriconi F (2003) Resti archeologici e analisi delle stratigrafie murarie. In: Barisi I, Fagiolo M, Madonna ML (eds) Villa d'Este. De Luca Editori d'Arte, Roma, pp 26–27

Conservation and Restoration Methods Applied to Ancient Ruins

58

Laura Pecchioli

“Raddoppiare l’area scoperta significherà offrire ai visitatori dell’Esposizione Universale una visione ben chiara e di indiscutibile interesse archeologico ed estetico, di una città romana imperiale che completa in modo meraviglioso il volto di Roma antica. E’ infatti Ostia che ci dà la cornice, l’inquadratura indispensabile per risentire e rivedere il cittadino romano intento alle sue occupazioni di ogni giorno. E’ Ostia che ci prospetta le questioni più attuali di urbanistica e di edilizia urbana e talvolta ce ne presenta le soluzioni geniali. E’ Ostia che ci fa conoscere le origini romane di molti motivi architettonici e decorativi, erroneamente ritenuti originali di formazioni artistiche postromane. E’ Ostia che ci dà quel che neppure Pompei od Ercolano possono darci, perché in parte lontane in parte anteriori al dinamismo sociale ed urbanistico dei tre secoli dell’Impero, in cui invece la vita moderna ritrova le più profonde. . ..?” (Guido Calza dalla rivista l’Urbe)

Contents 58.1

58.2

A Short Introduction to Cultural Heritage Regulatory Framework . . . . . . . . . . . . . . 58.1.1 Cultural Debate at the End of the Nineteenth Century . . . . . . . . . . . . . . . . . . 58.1.2 Charters of the Restoration (1931–1939) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A Brief History of Ostia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.2.1 Ostia: Source of Construction Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.2.2 Geological Setting and Natural Disasters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1724 1724 1725 1726 1727 1728

L. Pecchioli (*) Humboldt Universität, Winckelmann-Institut, Klassische Archäologie, Ostia Forum Projekt, Berlin, Germany Technische Universität Wien, Baugeschichte und Bauforschung, Berlin, Germany e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_58

1723

1724

L. Pecchioli

58.3

1728 1730 1731 1733 1735 1738 1740 1741

First Excavations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.1 Stratigraphic Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.4 Archive Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.5 Preservation and Valorization in Ostia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.5.1 Liberation of the Monument . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.6 The Preparation to the Universal Exposition in Rome (E42) . . . . . . . . . . . . . . . . . . . . 58.7 Critical Analysis and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

The ancient Roman city of Ostia is one of the most significant examples in the chronicle of Italian archaeological restoration, in the first half of the last century. It allows to study ancient construction techniques and various approaches of restoration and conservation on an extensive citywide scale. In the field of cultural heritage, the choices of preservation and anastylosis applied in Ostia are still relevant topics of discussion even today. The choices made have partly polluted the interpretation of the ancient masonry building. In particular the preparation phase of the International Exhibition in 1942 has contributed to a huge loss of information, especially from the archaeological point of view due to an excavation conducted without a scientific method. Ostia shows widespread methodological experimentalism of protection and valorization of an archaeological site, well suited for being analyzed with a didactic approach. Keywords

Restoration and conservation work · Anastylosis · Mimesis · Carta del Restauro · Conservation · Athens Charter · Cultural heritage

58.1

A Short Introduction to Cultural Heritage Regulatory Framework

58.1.1 Cultural Debate at the End of the Nineteenth Century An important cultural discussion characterizes the period at the end of the nineteenth century: in 1882 new guidelines are promoted by the ministry. The content outlines a procedure based on preliminary studies that would promote knowledge of the monuments and limit improvisation and unnecessary rebuilding. If rebuilding is performed, it is recommended to rebuild “such as it was” to maintain the character of the monument. The guidelines put forward a methodological approach that considers essential the historical and constructive characteristics including the causes and effects of the damage on the monument. The priority is given to the true character of the monument. Alongside a criticalphilological process identification and transmission of the building’s identity already suggested in that official document, it led off also the concept of authenticity as a determining feature for the recognition of the historical value of the

58

Conservation and Restoration Methods Applied to Ancient Ruins

1725

monument. This approach required a level of maintenance standards which was not possible at that time. An attempt to put in practice the guidelines was made by Fiorelli who, after having written the Circolare of the Decreto, established a laboratory for the conservation of materials. Regional Offices for the Conservation of Monuments were created having conflicts with Provincial Commissions and the Inspectors of Excavations’ management. Those conflicts led to a low interpretative quality of both restorations and intervention techniques, indifferent to the constructive peculiarities. With respect to the discussion of this period, the philological restoration approach, generally attributed to Camillo Boito (1836–1914), is notable and will shape the future discourse. It advocates the absolute respect and distinguishability of the additions as expressed in the Fourth Congress of Engineers and Architects held in Rome in 1883 [1]. The context of Ostia can show an interesting overview of the various conservation and restoration interventions, because they were performed precisely in those years of methodological debate in Italy. Depending on the type of conservation approach, in the intervention the difference between ancient and modern construction techniques can be difficult to see without careful investigations and knowledge of the materials. These choices are still today the topic of debates in symposiums to implement charters on restoration.

58.1.2 Charters of the Restoration (1931–1939) The definition of the first charters on restoration opened a relevant debate. The use of new technologies and materials “foreign” to the monument determined the need of defining those criteria of intervention. The Athens Charter, the first charter of the restoration written in 1931 by the International Conference of Architects gathered in Athens, wishes for a philological restoration approach and allows the use of modern materials for consolidation and anastylosis. The latter indicates specific types of interventions on monuments preserved in a ruined condition in order to, wherever possible, reinstate any original fragments that may be recovered. Thus, philological restoration is officially endorsed: “monuments must be rather consolidated than repaired, rather repaired than restored; all parts of a building must be respected, even those added during its history; if a new part is to be added to the building, it must be differentiated in terms of materials and characters, but without altering the overall aspect of the monument.” The IV Chapter is dedicated to the archaeological restoration, where one can find the criteria to protect the structures from an excavation. Restoration projects must be subject to informed criticism to avoid mistakes that lead to the loss of the historic character and value of monuments. The Superior Council for Antiquities and Fine Arts issued the Carta del Restauro (1932) the first official directive of the Italian State on restoration. It approves the use of modern technologies for consolidation, dictates limits on reconstructions, and seeks to give more attention to conservation than restoration. Maintenance and

1726

L. Pecchioli

consolidation are essential practices. On the other hand, restoration intervention must neither be carried out on a hypothetical basis or nor dictated by needs of artistic-architectural unity. The completion of the original is disallowed; only anastylosis and the least integration work are allowed. In paragraphs 10 and 11, the argument is about the preservation of the remains in situ, a topic connected to the method of the liberation of the monument. It stated principles similar to those of the Athens Charter, but with the attitude expressed by Gustavo Giovannoni about a scientific restoration (the approach for all historic buildings, not just classical monuments). He encourages the use of traditional techniques and “primitive” materials as close as possible to the original. Giovannoni particularly emphasized the formerly discounted value of the “minor architecture” of historic urban centers and towns, which make an important contribution to the overall historic environment [2]. Gustavo Giovannoni in 1916 was an appointed member of the Superior Council of Antiquities and Fine Arts: this role of ministerial superconsultant allowed him to examine a large number of projects located at the intersection of problems and different scales spread throughout the country. From this observatory he has been able to range from the themes of architectural restoration to those more related to the urban dimension, from the questions of the historian to those of the designer for 40 years. His theories influenced Italian and European legislation in conservation and planning, including also the Carta del Restauro [3, 4]. Giovannoni claims that in every choice, it is necessary to exploit all the most modern engineering sciences in order to achieve scientific restoration interventions. He encourages the full treatment of maintenance and consolidation confirming the respect for every single part of the monument and its environment. In particular, Giovannoni declares the necessity to investigate the monument and its history, before rebuilding it, criticizing the approach adopted in Ostia. Still only in the case of consolidation in unstable structures and reinforcements using new materials, especially reinforced concrete, is justified. The latter must be thought of in artistic harmony with the monument [5]. In 1939 two laws on the safeguard were promulgated. Law no. 1089 focuses on “things of art” (cose d’arte) thus including only significant heritage from the aesthetic point of view and only heritage consisting of material objects. At the same time, Law no.1497 mentions the environmental protection of the natural beauties. The Second World War, with all the physical destruction to the European architectural heritage, brought back the problem of architectural restoration. At this juncture, also due to the psychological effects of the desire to erase the destruction of war, the practice of restoration shifted often toward reconstructing the pre-existing, even at the risk of committing real historical falsehoods.

58.2

A Brief History of Ostia

As the harbor of Rome, Ostia was located on the ancient shoreline at the mouth of the Tiber River. During the Imperial Age, the city grew becoming an important commercial center [6]. The gain of the importance, from the Augustan period onwards, is

58

Conservation and Restoration Methods Applied to Ancient Ruins

1727

testified by the construction of several public monuments such as stocks and warehouses. Under Trajan and Hadrian, large civil and commercial quarters with several multistory insulae were built mainly in brick-faced concrete for outer walls and Opus mixtum for inner walls, brick-faced concrete with panels of Opus reticulatum [7–9]. After the years of the republic and the first centuries of the empire, a period of maximum splendor for the commercial activities that characterized the city, the beginning of the decline began in the middle of the third century AD with the political and economic crisis that hit the entire Roman empire. Following the decline of the Empire and despite a Renaissance, the town was progressively abandoned starting with the fifth century, followed by a slow and progressive depopulation of the urban area and the first reuse of materials and the first plundering of marble and furniture of buildings in Ostia began. In the case also a further critical element was represented by the growing importance of Porto during the empire of Constantine who took the name of Civitas Flavia Constantiniana Portuensis and obtained full municipal autonomy no longer depending on Ostia. A relative rejuvenation, to contrast the Saracens invasions, occurred only in the ninth century with the birthing of Gregoriopoli by Pope Gregorio IV in the area adjacent to the ancient town.

58.2.1 Ostia: Source of Construction Materials Ostia has provided materials for restoring itself and for the construction of the village (Borgo di Ostia) created to defend against pirate incursions. In the ninth century, the real plundering of Ostia’s monuments for the construction of other buildings began. In particular, numerous blocks of marble were taken and brought to Orvieto for the construction of the Duomo and others also to Pisa, where they were used for the construction and decoration of buildings inside Piazza dei Miracoli. All those events make it clear that over the centuries, the ruins of Ostia have been the subject of looting and plundering by invaders and defenders alike due to their presence at the mouth of the Tiber, point of landing for the ascent of the river to Rome. With this continuous passage and the continuous attendance of the area, the memory of Ostia was never lost and the plundering and the reuse of materials, began immediately. The testimony of the passage to Ostia by King Richard the Lionheart in August 1190 is indicative: “intravit Tyberim; ad cuius introitum est turris sed solitaria. Sunt et ibi ruinae maximae muro rum antiquorum. . . Vicesima sexta die Augusti transivit rex per quoddam nemus quater viginti miliaria,” in a landscape characterized by woods remains of roads (Via Severiana) and ruins of buildings from which material for construction was obtained. With the Renaissance started, the search for treasures of ancient art and the collection of inscriptions and Rome obviously had a privileged access to Ostiense material thanks to the Tiber as a means of communication. In 1598 under Clement VIII, the right to quarry marble from the ruins of Ostia was sanctioned for use in the Fabbrica di S. Pietro and for building works in the city. A large block of African

1728

L. Pecchioli

marble was used as a base for the statue of St. Peter on the top of the Trajan Column, while other marbles for the Basilica di San Giovanni in Laterano.

58.2.2 Geological Setting and Natural Disasters Many ancient cities of the Mediterranean are affected by natural disasters [10, 11]. The geological-geomorphological setting of Ostia indicates that it was developed along the alluvial plain of the Tiber River [12, 13], whose soft sedimentary package has played a key role in the construction and the preservation of the ancient Roman city [14] which was often repaired. From the beginning until the end of the nineteenth century the construction activities of Ostia, built on a sandy soil, was a difficult task, due to regular visits by natural disasters, such as flooding, fluctuations of groundwater, and earthquakes. In reference to the floods, in 359 A.D. Ammiano Marcellino testifies to a sacrifice made by the urban prefect Tertullo in honor of the Dioscuri inside a temple dedicated to them “in order to calm the waters of the sea,” since it allows us to assess how some monumental buildings were still used as for the multistory buildings. Today we can observe a different raising of ground levels, the position close to the Tiber has probably led to improve flood protection. Perhaps the aim of this choice may have been establishing a more solid basis for new buildings and/or as a flood protection [15]. In collaboration with INGV (Istituto Nazionale di Geofisica e Vulcanologia) of Rome, in 2017 a seismic survey was carried out placing sensors in some points near interesting damage case studies of the ancient city of Ostia and in Portus. This confirmed our estimates of the thickness of alluvial sediments (about 30 m). Through a crosschecking on the geomorphological literature, one has been identified a fault in the direction of W-SW/E-NE that was already active in the past. The comparison with the results of a mechanical analysis allowed us to support the hypothesis [16]. The fault would be located on the northern boundary of the city (Fig. 58.1). It may therefore appear that their activity has been accompanied by seismogenic phenomena which are responsible for the damage to the buildings in Ostia and caused local earthquakes [17].

58.3

First Excavations

A historical framework on methodological development in archaeological restoration is required, to better understand the culture of the time and the choices adopted. After a more or less long period of careless for the ruins of Ostia, a new interest begins in the eighteenth century with several excavations conducted within the urban area. In 1788 excavations, conducted by the Scottish painter Gavin Hamilton in the area of the Terme di Porta Marina, brought to the European antique market numerous pieces of art from Ostia that ended up in the English, French, and Russian collections. Unlike explorations for the antique market or private collections (Hamilton, Albani, Montanari, Fagan) and the papal excavations of the 1800s (Fea, Petrini, Bell, Visconti), the earthworks are still aimed at discovering works of art.

58

Conservation and Restoration Methods Applied to Ancient Ruins

(B)

(from Ciotoli et al., 2015)

1729

N

N

horizontal stress directions from collapsed structures

Inferred fault

Trajan’s Port

Beach ridge >3000 BP Roman epoch Post-Roman

Tiber River

Fiumicino channel

1700 BP shoreline Margin of the pre-reclamation Ostia lagoon

c

A

(from Bellotti et al., 2011)

“fiume morto”

B Ty rrh en

C

Topographic relief from Amenduini, 1884.

paleolagoon

isola sacra

SO2

Ostia antica

Borehole

ia n

Salomon et al., 2014 Salomon et al., 2016 Bellotti et al., 2011 Vittori et al., 2014

ing en ep de

a Se

I

“Fiume morto” MO3

LOA1

MO1

b

MO2 SO6

a

1 km

(A)

Fiumara Grande channel

ing

en deep

ISF1

PO2 PO1

II

b SO1

a

Fig. 58.1 (a) Structural-geomorphological map of the investigated area, showing the location of the inferred faults. The anomalous deepening at increasing distance from the mouth of the Fiumara Grande channel is highlighted; (b) structural scheme showing the fault system forming the halfgraben of the Tiber delta, characterized by the progressive southeastern migration (a, b, c) of the main fault segment (from [17]). (Source: F. Marra (INGV))

Only since the beginning of the nineteenth century, the archaeological excavations promoted by Pope Pio VII really began to take back to the light the remains of the ancient city. In the beginning the intention was also that of repairing and maintaining the archaeological structures of new excavation and those already exposed. Despite the aim, the first approaches were poorly documented and created serious problems to recognize the ancient masonries. In 1865 with the instructions of the Ministro della Pubblica Istruzione del Regno d’Italia, guidelines were issued (Istruzioni per gli Scavi di Antichità) attempting to indicate a method of excavation and protection of the ruins. It was geared toward the excavation of entire cities, extensive exploration with the relocation of collapsed elements, and protective measures taken over immediately after digging. Those regulatory provisions were at least partly implemented from 1871 with the Nuovo Governo Italiano of Ostia [18]. Rodolfo Lanciani was appointed director of excavations and professor of ancient topography at the University of Rome in 1878 [19, 20]. His research, although far from the currently accepted methods, had the merit of promoting systematic investigations aimed at the topographical understanding and to get the interest of the scientific world on Ostia. The investigations of the last 30 years of the century here briefly described were also the first organized restoration and conservation works. From 1911 the restoration and conservation activities began to be performed on a regular basis, and maintenance was applied widely. The first systematic excavation using an innovative management happened under the direction of Dante Vaglieri in charge of the excavations of Ostia from 1907 to 1913. In 1908 Vaglieri acquires a

1730

L. Pecchioli

Fig. 58.2 Decauville project, for the removal of a large amount of earth from the excavations and for the transport to reinforce the embankment of the Tiber, early twentieth century (Arch.fot.Ostia, n B2585)

system based on the railroad, Decauville project, a demountable solution for the removal of a large amount of earth from the excavations and for the transport to reinforce the embankment of the Tiber (Fig. 58.2). He began to discuss the relative issues of documentation and conservation of the ruins brought to light and with the desire to provide recommendations for possible attempts at reconstruction (Relazioni Quindicinali, 10.09.1910). His detailed documentation, analytical description, and stratigraphic reading, as well as the application of the method of anastylosis and recomposition or partial reconstruction, represent an important milestone in the preservation program in that time. Nowadays, thanks to the prolonged excavation and restoration works, much of the remains stand in relatively good conditions and represent a unique opportunity to study Roman construction techniques on an extensive and citywide scale.

58.3.1 Stratigraphic Method In 1901 Giacomo Boni publishes the guidelines of the new method of stratigraphic excavation in the magazine the New Anthology. He was a supporter of the need to preserve the authenticity and historical character of the monument and kept deep relations with the English cultural world [21]. The method of stratigraphic

58

Conservation and Restoration Methods Applied to Ancient Ruins

1731

excavation formulated by Boni was applied in Ostia principally by Vaglieri, also if partially, in numerous essays aimed at studying the Republican age. Later Guido Calza adopted it only for the sondages, supporting the preservation of all the structural and decorative elements. Unluckily the excavators would have removed many late-antique and early-medieval walls, partly due to their vulnerable state of conservation but mainly facilitated the recovery of the Imperial Ostia. Calza preserved at least the most significant testimonies of the Republican age and documents them [22, 23]. The stratigraphic interpretation has been unfortunately partially compromised by the alterations suffered by the ruins during the execution of the excavations and reconstructive works (Fig. 58.3). Over the years of the excavation, the stratigraphic method was more and more neglected. In 1936 the excavation method did not have any stratigraphic value and was reduced exclusively to the techniques of reallocation of architectural fragments. It is not possible to reconstruct exactly the scenario of the collapse that occurred; a structural analysis of various collapses was not foreseen in the restoration program, but the photographic and archival sources fortunately allow to recover some traces of the collapse dynamic [16].

58.4

Archive Documentation

The first quarter of the twentieth century is documented in Giornali di Scavo and Relazioni Quindicinali. Especially from 1907 to 1924, those are compiled by Raffaele Finelli, who significantly contributed to the management of the construction sites and to the interpretation of the findings. Rich and refined documentation of designs and details were produced by Italo Gismondi. Their works complete and enrich the archive documentation often resulting only in few photographs in certain periods of time. By reading the notes, descriptions, and drafts, we can know and analyze which materials, methodologies, and the situation in situ at the time of discovery [24, 25]. Although some monuments lack an exhaustive documentation about their excavation, the photographic material is a source and of extreme importance to define the chronology of the various vicissitudes of the monuments. Vaglieri had also the idea of inviting the Battalion Specialists (Battaglione Specialisti) of the Genius to perform the topographic survey of Ostia from the balloon brought in scale 1:2500 [26], a novel surveying approach in archaeology (at the time) to document the conservation status of Ostia in its ensemble (Fig. 58.4). Information on the quality of the materials, craftsmen, and construction sites can be searched in the abovementioned archives. Especially during the preparation of the international exposition of 1942 (E42, see also Sect. 58.6), the quality and the employment of the construction materials were not a priority. The choices influenced the preservation of the archaeological site. Here we discuss only some notes from Relazioni Quindicinali about the quality of the used mortars. A kind of aerial lime mortar is one of the most commonly used binders in Ostia, because with it, it is

1732

L. Pecchioli

Fig. 58.3 Caseggiato degli Aurighi, Reg. III, beginning of excavation (Arch. fot. Ostia, no. B2615)

possible to make a hydraulic mortar by adding incoherent volcanic products. In order to improve the physical and chemical qualities of the final product, an aggregate (normally sand) is added. Normal pit lime is obtained by quenching quicklime in clods and has a pasty and creamy consistency to the touch, especially if its purity is high. Generally, the aggregate (sand) originates from river beds, beach deposits, and sandy outcrops, but in the region of Rome and Ostia, the same incoherent volcanic deposits are often used as an aggregate. About the operating procedures, the aerial lime mortar does not always seem produced with the optimal maturation times. Often rainwater is used, replaced later even with seawater or with water from shallow water wells. One or more parts in volume of red pozzolan with sand (harena) are often substituted using sand dug directly from the soil of Ostia also for economic reasons. The latter procedure, introduced by Finelli in 1908, has been applied also in the following years, especially in the financially difficult periods to have canonical hydraulic mortar. These varieties have lower – hydraulic – resistance and a longer setting time, due to the addition of sand, and are thus used in restoration work, where high static performance is not required. The use of traditional mortars in lime and pozzolan is suggested, in addition to direct observation, by the descriptions of the excavation diaries, from which it is obtained that the lime used is extinguished in trenches and, if necessary, used immediately after the shutdown without any maturing (the mortars

58

Conservation and Restoration Methods Applied to Ancient Ruins

1733

Fig. 58.4 Ostia, 1918. Oblique view of the sector north of the Decumanus between the Capitolium and the Great Horrea (Arch. fot. Ostia, no. B2430)

consisted mainly of lime and pozzolan and required quantities that one can hypothesize mixed in a volume ratio of 1:3). Another resolution adopted for short periods is also the addition to the mixture of crushed and ground rubble. Many of the mortars have brought on a characteristic pinkish tone, indicating the greater use of the red pozzolanic aggregate in a finer grain size (almost a powder). The usage of cement was limited to the consolidation and restoration of the decorative apparatus, and the use of bastard mortar was not excluded [27].

58.5

Preservation and Valorization in Ostia

After the debate at the end of the nineteenth century, different techniques of restoration to put in practice the ideas discussed succeed one another such as methods used to remark the modern interventions as, e.g., using the brick wall appliance placed irregularly at 45
, a method that is called sawtooth (addentellato), left in sight to suggest the continuation of only partially preserved wall sections. Another technique often used, is the so-called subasquare (sottoquadro), a method useful to place the restoration surface not coplanar to the old but slightly backward compared to it. Between 1877 and 1889, the research of Lanciani, even if based on the method of the preventive archaeological survey and earthworks (methods applied therefore

1734

L. Pecchioli

without a scientific criteria), promoted systematic investigations in order to understand the topography and the functioning of the ancient city that spurred scientific interest on Ostia. In 1882, the same ministry tries to encourage the development of a preliminary phase in the restoration, although still not very critical-philological. The approach should have taken into account the historical-constructive characteristics and respects the concepts of identity and authenticity of the monument. In the final 30 years of the nineteenth century, the restoration works are conceived above all as reconstructive operations to restore the compromised static conditions and as an attempt to reconfirm the legibility/comprehension of the lost architectural structures: consolidation, restoration, raising of fallen parts, detachment, or restoration of mosaics or paintings is held out. Vaglieri championed the method of anastylosis and recomposition or partial reconstruction. He recovered ancient bricks to complete and integrate the ruins with masonry systems of old bricks recovered without any criterion of differentiation [28]. The peculiar use in restorations of the almost exclusive use of ancient bricks, recovered during excavation operations, has continued to be adopted for a long time due to the great availability on the site (Figs. 58.5 and 58.6). This approach called mimesis can be seen, for example, in the Terme of Neptune, Theatre (there are besides several in situ consolidations of the most significant collapses relevant to the higher levels). His desire was to present an overall vision of Ostia at its most emerging stage, an imperial city even using a methodology based on an unscientific approach. We move away from the approaches of Rosa, Lanciani, and Moreschi based on sawtooth bricks, bricks put in place with edges in the visible face. Ostia has been abandoned for about 10 years until the beginning of the twentieth century, when De Angelis established to protect the top wall solutions called Bauletto in cocciopesto against rainwater and vegetation (Fig. 58.8). The solution is adopted as a protection solution in the shape of a trunk, obtained by repeated beating of a traditional mortar of lime and pozzolan, with fragments of bricks of different sizes. There was also a tendency to “regularize the surfaces” by proceeding with arbitrary demolitions, and, depending on the situation of the ruin, in some cases, the masonry was brought to life, in order to proceed with the laying of the cocciopesto. This practice avoids costly interventions of restoration and consolidation and is adopted until the beginning of the Second World War. Vaglieri is responsible for the first systematic excavation activity: his documentation activity is scrupulous, even if the attempts at stratigraphic reading are aimed at documenting above all the Republican age. His works program is indicative and is divided into three simple points: to complete the excavation of the buildings previously not completely exposed, taking care of the conservation of the ruins already excavated; to join the individual groups of ruins; and to make the excavations in depth and clarify the progress of the history of Ostia. We move away from the approaches of Rosa, Lanciani, and Moreschi based on sawtooth clamping, bricks with edges on their faces. In restorations the almost exclusive use of ancient bricks, recovered during excavation operations, without the differentiation expected in the modern approach, continues to be adopted for a

58

Conservation and Restoration Methods Applied to Ancient Ruins

1735

Fig. 58.5 Terme del Foro, 1928. Some phases of the relocation of a collapse (divided into several parts) (Arch.fot.Ostia, no. C1260)

long time due to their great availability on the site. In the last quarter of the nineteenth century, the restoration works are designed primarily as reconstructive operations to restore the static conditions compromised and as an attempt to reconfirm a readability/understanding of lost architectural structures.

58.5.1 Liberation of the Monument In the first decade of the twentieth century, the figures which play a relevant role in the excavations of Ostia for the Universal Exhibition are Guido Calza and two

1736

L. Pecchioli

Fig. 58.6 (a) Horrea Epagathiana, 1941. The collapsed south wall. It will then be relocated to its original position (Istituto Nazionale di Archeologia e Storia dell’Arte, Fondo Ordinario, n. 5239). (b) Photo of the recomposed wall (2018)

58

Conservation and Restoration Methods Applied to Ancient Ruins

1737

collaborators as Italo Gismondi and Raffaele Finelli. He provides works of consolidation, restoration, raising of fallen pieces, detachment, or putting back in place of mosaics or paintings, all that serves to present in the best possible form the monumentality of Ostia. In these years, in order to facilitate understanding, recompositions have been adopted, and curtains of old bricks recovered without any criteria of distinction have been used. Vaglieri’s approach continues in the integrations without distinction, but slowly an approach with splintering of the edges or with parallel engravings, often in conjunction with a slight retreat from the edge of the ancient curtains, is also common. Between 1907 and 1924 according to the state of conservation of the ruins, there is an employment of ancient materials in respect with the original construction technique. In the first quarter of the twentieth century, the excavations (1900–1925) are based on the practice of isolating a monument, an approach that has been strongly supported also by Modern Urbanism from the second half of the nineteenth century. Archaeology is interpreted as a practice of liberation method, and the usage of the trench represents the methodology: in situ the recomposition is carried out and the collapses are raised (Fig. 58.4). The excavation heap is used as temporary scaffolding, a technique that allows to proceed with immediate relocations and a big saving of means, money, and materials. Liberation means the rapid evacuation of waste materials and the rapid discovery of what is of most interest. The trench is used in Ostia in particular in the necropolis. In addition, the excavations have to reach the archaeological plan, defined as the level of the ancient roadway or the floors of the earthly environments. Calza claims the scientific dignity of militant archaeology in L’archeologia della zappa e del piccone [22]. He supports the recovery of all the structural and decorative elements, favoring an imperial period, preserving, and documenting only the most significant testimonials of the republican age. Late antiquity and the early Middle Ages are not considered by Calza to be worthy of being preserved. With Guido Calza also the practice of mimesis starts, performed without distinction criteria and applied in cases of reparation. The approach is assumed without any documentation regarding their state of discovery. In 1912, after having completed his studies, Calza entered the Amministrazione Antichità e Belle Arti as an inspector under Vaglieri who since 1906 had been in charge of the excavations in Ostia and since 1908 with the title of director. It was not given to Vaglieri to see his goal realized, the resurrection of the ancient port of Rome. Calza judges the adoption of the cocciopesto cases from previous interventions as a strong aesthetic impact and the reason why he initially proposed leaving the tops of the walls unprotected and eventually covering them with grassy layers, in an attempt to reduce the process of breaking up the mortar. The daily observation in the manifestation of the processes of degradation, however, leads him to think again, namely, how to resort to the protection of the ridges of walls to cover with mortar and elements of tuff and brick. This practice was adopted until the beginning of the Second World War. Many of these mortars, like most of those

1738

L. Pecchioli

packaged in E42 works, have taken on a characteristic pinkish tone, favored by the greater use of the red pozzolanic aggregate in a finer grain size (almost a powder).

58.6

The Preparation to the Universal Exposition in Rome (E42)

Until the middle 1930s, the cult of the Roman Empire had not been included in a well-structured political program. But then the reference to the imperial Rome of antiquity, and to colonial expansion, became an ideal of the future, dominated by a fascist homo novus bringing the triumph of values and norms of antiquity [29, 30]. The Universal Exposition, E42, becomes an ideal event to affirm these ideas and to give prominence to the regime. Ostia in particular could meet those expectations and could show the greatness the regime wanted to claim for itself. This interest led to unscrupulous excavations both in necropolises and in historical centers. In Calza’s mimesis practice, one can note how an important role was played by the use of modern handmade bricks (26  13  3), both full and perforated, according to destinations and needs. They were more resistant to stress, less fragile, and more porous, almost comparable to the performance of old bricks. Their use finds application in the integration of the columns of the frigidarium of the Terme del Foro, for example, as in the lintels of the arches of Casa di Diana along Via dei Balconi. Modern bricks continue to be used alongside the ancient ones recovered in the phases of excavation flush with the ancient curtains [27, 31]. The Universal Exposition in Rome (E42) is an opportunity to rediscover the ancient Roman city. Extensive excavations brought to the surface long forgotten parts of the ancient city that were accompanied by a series of interventions in archaeological sites and liberation of the monuments [30]. More than 600,000 m3of earth were removed (Fig. 58.7) that had reached a height of 4–12 m above the ancient street level. Needless to say that much information was not recorded during these 5 years [32, 33]. In the 20 years of fascism, the archaeological practice begins when the work is finished, giving it a marginal scientific role. In a few years, large financial resources were invested, and between 1938 and 1942 the excavated area doubled to 34 hectares. Most of the funds were allocated to the restoration and archaeological arrangement of the ruins and to the excavation works. Through the observation of the photographic documentation, it can be seen that the extracted material from the excavations is often used as a temporary scaffolding for the immediate relocation of the collapses and reconstruction of the missing parts (Fig. 58.4). The adoption of this solution saves time and money but modifies the original stratigraphy and loses unique information. After the Second World War, the excavations underwent a period of stagnation, while the large number of restorations of the structures brought to light in previous years continued. It was only from the 1960s onwards that excavations resumed, this time limited to individual buildings and in many cases specifically dedicated to the discovery of layers from the Republican period of the city.

58

Conservation and Restoration Methods Applied to Ancient Ruins

1739

Fig. 58.7 Cardo Maximus, 1939–1940. Decauville project. Excavations during the Universal Exposition of Rome (Arch.fot.Ostia, n. B2946)

Fig. 58.8 Terme del Foro. 2018: (a) The difficulty of reading the ancient but also the various subsequent choices of restoration, integration, and consolidation; (b) Example of protection “a Bauletto” with restoration work on the masonry of 1924 and 1964

1740

58.7

L. Pecchioli

Critical Analysis and Conclusions

The ruins of Ostia have undergone many restoration and conservation operations; their extent varies depending on the cultural attitudes of the management and the levels of survival of the original buildings. The excavations in the nineteenth and twentieth centuries mirror the urban and architectural development of the imperial Roman city, although the approach and the methods of excavation and restoration are questionable. The stratigraphic reading method was abandoned by Calza due to the difficulty to apply it to the monumental topographical excavation of large ancient urban sites. The political event of the International Exhibition E42 has been the occasion to have the means for a great construction site in Ostia. This period coincides with an important phase of development for the recommendations on conservation and restoration. The earthwork is still unluckily a practice used in the archaeological campaign. The divergence between workers of companies and those employed in the technical office does not allow a timely check of the enormous amount of work. In this 4-year period, due to the haste to bring to light as many structures as possible, for the needs of fascist propaganda, a large excavation has been carried out in which the level of the second century AD (of the Hadrian age) was frenetically reached, without adequate documentation of the layers. In this way many data relating to the late ancient Ostia were totally lost, and also many restorations and reconstructions were carried out in a summary and often a false way. In Ostia one can find a rigorous application of the guiding principles of minimum intervention and distinctiveness and but also the total rejection of these principles and even original reinterpretations in a single place. In many cases, the surviving structures were recomposed and redesigned in order to restore the lost monumentality and to transform a city of rubble into conditions of comprehensibility. The reuse of ancient materials has led to mimetic results that today are not easily recognizable and often misunderstood with ancient transformations and restorations. The critical interpretation on the choices and on the development of the preservation practice in cultural heritage carried out in the last century requires an investigation on the impact of past interventions on the current possibilities of study of archaeological ruins (loss of stratigraphic information, difficulty in interpretation, and identifying for the mimetic integrations and of comprehensibility) (Fig. 58.6a, b). In spite of these problems, the reconstruction of methodologies offers the possibility of testing the outcome of the interventions adopted in those years, revealing their merits and defects. A detailed photo documentation allows us to complete various time frames of the ancient Roman city, in order to trace the choices and the restoration criteria [34]. The reconstruction based on the immediate recomposition of the collapse fragments of the large complexes has not been so oft documented. Guido Calza argues that the city of Ostia must be dug with a systematic exploration, and that in order to revive the imperial city in its original monumentality, it is necessary to use all the structural and decorative elements available from the layers of collapse and

58

Conservation and Restoration Methods Applied to Ancient Ruins

1741

abandonment. The restoration work was mainly intended as reconstructive operations aimed at restoring compromised static conditions and a readability of lost monuments [35]. The reconstruction is defined as the total or to a considerable percentage recreation of a vanished monument using also new materials. Since 1924 there has been a considerable lack of documentation on the excavation activity (after the retirement of Finelli). In the following decades, thanks to the increasing attention to the archaeological site, part of the resources was periodically used to clean and preserve previously discovered buildings. In archaeological areas, accepting to measure oneself against the lost ancient building, its functioning, and its materials and to promote a compatible and contextual valorization is the right conservation strategy. The main difficulties of protection and management of an archaeological area on an urban scale, such as Ostia, are due in the first place to the lack of a reliable knowledge of the state of conservation of individual archaeological architectures. It is necessary to initiate planned conservation actions through inspection and monitoring activities, adopting an operational tool that takes into account the specificity of the architecture of Ostia, in its twofold technical-constructive and historicalconservative peculiarity. The analysis of previous interventions and the study of the evolutionary dynamics of different pathologies are essential for a correct diagnosis of the state of health of the artifacts [36]. The possibility to examine the speed of progress of the degradation, through the comparison with the archive data, also allows to better define the conditions of risk, significantly reducing the current level of uncertainty. Acknowledgment I am grateful to Soprintendenza Archeologica di Ostia for the collaboration and to Gerda Henkel Stiftung for the backing, without which the development of this research would not have been possible.

References 1. Boito C (1885) I nostri vecchi monumenti. Conservare o restaurare? In: Nuova antologia di scienze, lettere ed arti. Jg. 21 ¼ Bd. 87 ¼ Serie 3, Bd. 3, 1886, ZDB-ID 211166-4, S. 480–506 2. Gunzburger Makas E, Stubbs JH (2011) Architectural conservation in Europe and the Americas. Wiley & Sons eds 3. Giovannoni G (1932) Cronaca. La conferenza Internazionale di Atene per il restauro dei Monumenti. Bollettino d’Arte 4. Giovannoni G (1913) Restauri di Monumenti. Conferenza di Gustavo Giovannoni, Bollettino d’Arte del Ministero della Pubblica Istruzione. Ed. Roma. http://www.bollettinodarte. beniculturali.it/opencms/export/BollettinoArteIt/sito-BollettinoArteIt/Contributi/Editoria/ BollettinoArte/Fascicoli/Fascicoli-Serie-I/visualizza_asset.html_1695903579.html 5. Dezzi Bardeschi C (2007) Archaology and conservation – theories, methodologies and field practices. Maggioli Editore, Milano 6. Paschetto L (1912) Ostia. Colonia Romana, Roma 7. DeLaine J, Wilkinson D (1999) The Reading Ostia project: excavation and survey in Insula I, IV, Ostia. In: S. Mols, C. van der Laan (eds) Mededelingen van het Nederlands Instituut te Rome: AntiquityNederlands Instituut te Rome, 58, 19

1742

L. Pecchioli

8. DeLaine J (2002) Building activity in Ostia in the second century AD. In: Bruun C, Gallina Zevi A (eds) Ostia e Portus: Nelle Loro Relazioni con Roma. Institutum Romanum Finlandiae, Roma, pp 41–101 9. Pavolini C A survey of excavations and studies on Ostia (2004–2014). J Roman Stud 106(2016):199–236. https://doi.org/10.1017/S0075435816001015 10. Pecchioli L (2019) Analysis of the historical restoration/repair measures and the stratigraphicarchaeological studies after Natural Catastrophe Events in Ostia Antica. In: Daniel Schneller/ Guido Lassau (Hrsg.) Erdbeben, Feuer, Wasser und andere Katastrophen. Ihr Einfluss auf die Stadtentwicklung und Stadtgestalt im Spätmittelalter und in der Frühen Neuzeit. Beiträge der Tagung in Basel 1./2. Februar 2018. Bern. ISBN 978-3-03797-597-8. https://peristyle.ch 11. Guidoboni E (2010) Upside-down landscapes: seismicity and seismic disasters in Italy. In: Armiero M, Hall M (eds) Nature and history in Modern Italy. Ohio University Press, Athens, pp 33–55 12. Salomon F, Goiran J-P, Noirot B, Pleuger E, Bukowiecki E, Mazzini I, Carbonel P, Gadhoum A, Arnaud P, Keay S, Zampini S, Kay S, Raddi M, Ghelli A, Pellegrino A, Morelli C, Germoni P (2018) Geoarchaeology of the Roman port-city of Ostia: Fluvio-coastal mobility, urban development and resilience. Earth Sci Rev 177:265–283. https://doi.org/10.1016/j.earscirev.2017. 10.003 13. Bellotti P, Calderoni G, Di Rita F, D’Orefice M, D’Amico C, Esu D, Magri D, Preite Martinez M, Tortora P, Valeri P (2011) The Tiber river delta plain (central Italy): coastal evolution and implications for the ancient Ostia Roman settlement. The Holocene 21(7): 1105–1116. https://doi.org/10.1177/0959683611400464 14. Arnoldus-Huyzendveld A, Paroli L (1995) Alcune considerazioni sullo sviluppo storico dell’ansa del Tevere presso Ostia. Archeologia Laziale 12:383–392 15. Jansen GCM (1995) Die Wasserversorgung und Kanalisation in Ostia Antica; Die ersten Ergebnisse. In: Mitteilungsheft der Frontinus-Gesellschaft 19:111–123 16. Pecchioli L, Cangi G, Marra F Evidence of seismic damages on ancient Roman buildings at Ostia: an arch mechanics approach. J Archaeol Sci Rep 21:117–127. https://doi.org/10.1016/j. jasrep.2018.07.006 17. Marra F, Milana G, Pecchioli L, Roselli P, Cangi G, Carlucci G, Famiani D, Mercuri A (2019) Historical faulting as the possible cause of earthquake damages in ancient Ostia (Rome, Italy): a combined structural, seismological and geological analysis. In: Pecchioli L, Panzera F and Poggi V (eds) Cultural heritage and Earthquakes: bridging the gap between geophysics, archaeoseismology and engineering. J Seismol. https://doi.org/10.1007/s10950-019-09844-z 18. Pallottino E (1994) Roma 1846–1878: restauro di monumenti antichi tra rappezzi mimetici e ricostruzioni semplificate. Ricerche di Storia dell’arte 52:69–71 19. Lanciani AR (1847–1929) Scavi Di Ostia. Coi tipi del Salviucci, Roma, 1881 20. Palombi D (2006) Rodolfo Lanciani: L’archeologia a Roma tra Ottocento e Novecento. “L’Erma” di Bretschneider, Roma 21. Boni G (1901) Il metodo negli scavi archeologici. “Nuova Antologia”, series IV, Vol. XCIV, 16 July, Rome, pp 312–322 22. Calza G (1926) L’archeologia della zappa e del piccone. Rassegna Italiana 102:3–15 23. Calza G (1936) Come si scava una città antica, da SAPERE – Ulrico Hoepli Ed., Anno II – Volume IV – n. 44, 31 ottobre – XV 24. Bedello M (2007) Tata: Ricostruire l’Antico Prima del Virtuale. Italo Gismondi. Un Architetto per l’Archeologia (1887–1974). Ministero per i Beni e le arttività Culturali, Soprintendenza archeologica di Roma, Archivio Storico a Palazzo Altemps 25. Ietto M (1995–96) Gli scavi di Ostia antica e l’attività di Guido Calza e Italo Gismondi nella formazione del dibattito culturale ed architettonico contemporaneo. Tesi di laurea, Università degli Studi di Roma La Sapienza 26. Shepherd EJ (2006) Il “Rilievo topofotografico di Ostia dal pallone” (1911). In: “AAerea” II, pp 15–38

58

Conservation and Restoration Methods Applied to Ancient Ruins

1743

27. Rinaldi E (2014) I restauri ostiensi di Vaglieri, In Bollettino di Archeologia Online, Direzione Generale per le Antichità, V,/2 28. Vaglieri D (1914) Ostia : cenni storici e guida/Dante Vaglieri; con cinque tavole e ventiquattro figure, E. Loescher & Co. (W. Regenberg), Roma 29. Visser R (2001) Da Atene a Roma, da Roma a Berlino. L’ Istituto di Studi Romani, il culto fascista della romanità e la „difesa dell’umanesimo“ di Giuseppe Bottai (1936–1943), 111–123. In: Zu den Forschungen über Antike und Altertumswissenschaften in der Zeit von Faschismus und Nationalsozialismus, Näf, B., Mandelbachtal/Cambridge: edition cicero 30. Kallis A (2014) The third Rome, 1922–1943: The making of the Fascist capital. https://doi.org/ 10.1057/9781137314031 31. Rinaldi E (2015) Ostia & Roma Archeologia e Restauro Architettura: Restauro e conservazione a Ostia nella prima metà del Novecento, Tesi di dottorato, ArcAdiA, Roma Tre (05/2012), 1–135 & RA restauro archeologico. Firenze, 47–67 32. D’Errico F, Pantò G (1985) La pratica e l’evoluzione del fare archeologia in Italia nell’esame di “Notizie scavi”, anni 1926–1943. In Archeologia medievale. Cultura materiale, insediamenti, territorio, 12 33. Olivanti P (1999) Il Caseggiato del Serapide e le Terme dei Sette Sapienti: scavo e restauro ad Ostia prima dei “grandi sterri” per l’Esposizione Universale del 1942. Mededeelingen van het Nederlandsch Historisch Instituut te Rome 58:11–14 34. Piccirilli C (1996) Consolidamento critico, premesse storico-strutturali, Saggio introduttivo di Carbonara Giovanni. Bonsignori Editore, Roma 35. Pecchioli L (2003) Problematiche connesse al restauro archeologico, The Cultural heritage protection in Italy. In: Mediterraneum, La Tutela dei Beni Culturali in Italia, Università degli Studi di Napoli„L‘Orientale“ – Facoltà di Studi Arabo-Islamici e del Mediterraneo, Fabio Maniscalco (a cura di), v.1, Massa Editore, Napoli, pp 61–63 36. Marino L (2009) Material for an atlas of pathologies in archaeological areas and ruined buildings. Alinea Editore, Firenze

Sustainable Conservation and Restoration of Historical Gardens

59

Rachele Manganelli Del Fa`, Alberto Casciani, Silvia Vettori, Oana A. Cuzman, Piero Tiano, Paola Rosa, and Cristiano Riminesi

Contents 59.1 59.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Architecture and Green . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.2.1 History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.2.2 Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3 Stone Furniture and Decay Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.1 Marble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.2 Sandstone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.4 Evaluation of Decay: Multidisciplinary Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.4.1 Laboratory Analytical Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.4.2 Biodeteriogen Investigation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.5 Sustainability and Minimal Intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1746 1748 1748 1749 1751 1752 1753 1753 1754 1755 1756 1759 1760

Abstract

A garden is a complexity of components with very different evolution and decay rates (cit.) (Dezzi Bardeschi M, Restauro: punto e da capo. Frammenti per una (impossibile) teoria, a cura di V. Locatelli, 8a Edizione. Franco Angeli, Milano, 2009). In historical gardens, artworks are constituted by heterogeneous materials such as natural and artificial stones that are subject to different deterioration. The

R. Manganelli Del Fà (*) · S. Vettori · O. A. Cuzman · P. Tiano · C. Riminesi Institute of Heritage Science – CNR, National Research Council, Florence, Italy e-mail: [email protected]; [email protected]; [email protected]; [email protected]; [email protected] A. Casciani Ditta Casciani Alberto Conservazione e Restauro, Florence, Italy P. Rosa Paola Rosa Conservazione e Restauro opere d’Arte, Florence, Italy © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_59

1745

R. Manganelli Del Fa` et al.

1746

main decay phenomena depend on the characteristics of the original material, the presence of metal structures of completion, and environmental factors, both natural and anthropogenic. The environment and minerals salts favor the proliferation of flora and microflora that are intimately related to the chemical composition and morphological structure, and the microbial communities with more biodiversity are those able to degrade the materials selectively. Other damages are due to accidents or vandalism. An interdisciplinary diagnostic approach is fundamental to understand the state of conservation of artworks outdoors, determine the causes of decay, and choose between restoration and maintenance intervention. Nonetheless, a sustainable choice of minimized intervention at a low impact on artworks and the environment is a priority. In this contribution, the research activity and the results obtained in several case studies are discussed.

59.1

Introduction

Garden art is an expression used for the first time by Fabio Colonna in 1499, in his work called “Hypnerotomachia Poliphili,” in which Poliphili travels in his dream to Citera Island and describes in great detail the beautiful luxuriant gardens in different allegoric scenes [1]. A garden is an environment in which vegetation is integrated with the architecture; understood as the result of the human intellect, it must be considered a work of art. In the garden, natural elements – such as trees, rocks, and water, also in artificial form – and architectural decorative elements are combined. The choice to use natural rather than architectural elements is closely linked to an era and a specific culture. Since the garden is designed by men, it has always been influenced by the aesthetic, social, and economic values of a particular culture or a specific historical moment [2]. Two components exist together in the garden, architecture and vegetation, and should be kept in balance with each other. The natural growing and evolution of vegetation and the biodegradation process affecting the surface of stone monuments can represent a risk for the conservation of the architectural parts. These problems can show themselves both from a macroscopic and microscopic point of view. In the scenario of the macroscopic aspect, the current conservation status of trees in many historical gardens is precarious, due to both their age and the effect of a rapidly changing climate. Moreover, the natural processes of senescence, various types of biotic and abiotic stresses, and extreme events associated with global warming (e.g., droughts and wind storms) have recently caused serious injuries to many old trees in recent decades. As a consequence, the physiological conditions and stability of many trees cannot completely avoid the risk of damage on stone artifacts or, even, to visitors. On the other hand, from a microscopic point of view, the effects of biodegradation due to lichens, algae, biofilms, etc. colonizing the surfaces of stone artifacts are the main issue.

59

Sustainable Conservation and Restoration of Historical Gardens

1747

For a long time, the Soprintendenza of Cultural Heritage have focused their interest mostly on the monuments in the strictest sense of the word, ignoring the naturalistic aspect. Their lack of awareness when considering gardens as artistic creations (for the variety of plants and the multiplicity of sculptures and architecture), in the past, has led to the destruction of these environments. Furthermore, the postwar urban planning policy that continued until a few decades ago had caused serious damage to the landscape, with the construction of new residential areas and extensive public works. The gardens built in hilly and coastal areas were absorbed in the cities, and they became part of the urban periphery, and a large portion of green spaces became urbanized. To avoid the indiscriminate use of the soil, the urban planning standard was born. This planning identifies and limits parts of the territory for a specific destination (residence, industry, and so on). The requirement to leave some areas to green areas, parks or gardens, has favored the exploitation of historic gardens and exposed them to risk, because historical gardens were no longer considered an artistic and cultural asset but only a public green area. Therefore, the final aim of the intervention of restoration will be to safeguard the original design and architecture of the garden by using innovative approaches and techniques to preserve biodiversity, cultural, and historical components. The proposed solutions must meet the future needs and balance both the vegetation and stone and architectural components, to develop best practices. These practices could include the substitution of single individuals or groups of trees or plants, which are no longer sustainable or suitable to the site also as a consequence of the current and expected climatic modifications, or substitute copies of stone manufacts as an extreme solution for preservation. The concept of protecting historical gardens in Italy, concerning the established principles of the restoration of architectural, pictorial, and sculptural assets, is relatively recent. Although this topic was already known and discussed in international circles, it reached a greater awareness only with the Carta di Firenze of 1981 [3]. On May 21, 1981, the ICOMOS-IFLA (International Committee of Historic Gardens) met in Florence and drew up a document called Carta di Firenze about preserving historic gardens. The interest in the enhancement and protection of historical gardens and parks has begun only in recent years, with the awareness that parks and gardens are an integral part of the historical-cultural and environmental heritage of our country. Managing historical gardens requires first dividing the whole area into homogeneous zones, which are independent of the traits and conservation status of garden areas. The division will follow general criteria such as geological and topographic characteristics; type of visitor access; and identification and nature of potential targets (monuments and/or buildings, tree and/or branch failure), historical and landscape relevance, etc. The obtained zoning must be compared with any previous zoning and inventory. The monitoring actions by enhanced technologies will complete the conservation status picture of the garden and is a prerequisite for proposing new restoration interventions.

R. Manganelli Del Fa` et al.

1748

59.2

Architecture and Green

59.2.1 History A garden is a complex, multi-material artifact: marbles, stones, and different metals are used in architecture and compositions of statues; mosaics and fountains with their complex hydraulic equipment and in the end, living matter such as trees, plants, and flowers [4]. Architecture has always been an important part of the garden during the fifteenth century, when it was conceived as an organization of different elements, mainly vegetal, independent of the house and intended as an extension of the structure of the building to the outside. The garden was intended as a link between the volume of the villa/home and the surrounding environment. In the following century, the garden was enriched with architectural forms. For the design of the gardens, the architects were inspired by the ancient Roman villa: statues, marble, caves, pergolas, water features, and the modeling of evergreen essences according to topiary art [5–7]. In the Renaissance garden, designers want to celebrate the victory of man over nature and the dominance of the intellect over natural events. Terraces, labyrinths, pergolas, etc. symbolically represent man’s planning operations on the territory, while rocks, artificial caves, etc. express the ability of the artist/architect to overcome nature and amaze the visitor. Caves and nymphaeums are typical elements of the garden of the sixteenth century. They are small habitats with a naturalistic aspect, designed for refreshments during summer days. These habitats were made of masonry and decorated with calcareous concretions, sponges, and shells in imitation of nature, and often they hosted water games to amuse visitors. In the design of a garden, water was fundamental: fountains, with the complex systems of water adduction and runoff, deeply influenced the location, the layout, and the development of the garden. In this regard, the gardens were built on the side of relief to bring water through a fall system. But the Renaissance garden is also characterized by initiatory paths and symbolic itineraries. Within the vegetation, it is often possible to find architectural and sculptural compositions, intended to compose a story. However, there are not only architectural compositions but vegetal ones as well, designed for hunting and delight such as labyrinths, cerchiate (arched arbors of bent tree branches) and ragnaie (groups of rows of trees, strung with nets to hunt birds). In the 1600s, especially in France, the Baroque garden developed [8]. The garden in this period reflects the taste of the clients (sovereign and nobles) and models Italian Renaissance gardens adjusting them to the grandeur of the nobility. The Baroque garden is built on a territorial scale: dimensions increase, and perspectives become complex and interact with the surrounding landscape, whereas the garden is enriched with increasingly complex sculptural apparatuses, fountains, and theatrical productions, with allusions to mythological and allegorical themes. In the following century, the Baroque style was still fashionable. The art of gardens which had enriched the Renaissance garden with amazing effects during

59

Sustainable Conservation and Restoration of Historical Gardens

1749

the previous century acquired a European style during the seventeenth century: the residences were still inspired by the Baroque trend, but gardens were enriched with naturalistic views, anticipating the pictorial style that was spreading throughout England. Meanwhile, in England, writers, philosophers, and men of culture condemn the artificiality of the formal garden as the result of man’s arrogant attitude towards nature, preferring a landscape composition inspired by the natural harmony of things. The architecture and decorative elements, such as sculptures and fountains, assume less importance in this period. Real-scale walk paths appear in gardens, the fountains and canals turn into ponds and streams, and trees and shrubs take on a naturalistic appearance. For the design of the garden, we are inspired by the works of contemporary painters such as N. Poussin, C. Lorrain, and S. Rosa, which represent tormented landscapes of central-southern Italy characterized by ruins, rocky spurs, and streams. In the so-called romantic park, there are often archaeological remains, real or artificial, and other elements such as folies and fabriques, small architectural objects, or small thematic buildings. The function of these buildings is to attract the visitor’s gaze and evoke places far away in time and space. In the nineteenth century, the event that most influenced the course of time, and human life was undoubtedly the Industrial Revolution. There were numerous changes: the abandonment of the countryside, city growth, the increase in mechanization, and the change in the relationship between man and nature. But in the nineteenth century, we are also witnessing the discovery and cultivation of new exotic species, which gave birth to a new garden aesthetic. A new aesthetic taste developed: the garden became pittoresco with particular attention to the search for contrasts and effects. Artists chose to use architectural elements such as ruins, chinoiserie, neo-Gothic or neoclassical elements, exotic elements like pagodas, Gothic tombs, columns, large amphorae, temples, and greenhouses used to cultivate particular plants. At the turn of the nineteenth and early twentieth century, the Liberty Garden developed following the fashion of extra-urban villas that started out as holiday resorts and then became refined symbols of the new bourgeoisie of the time. These are gardens where exotic species from other continents were cultivated, becoming places of leisure and entertainment.

59.2.2 Interaction As previously stated, in the garden natural and artificial elements coexist, and the degradation processes on stone furnishings are induced by the environment in which they are. All stone materials present in the monumental gardens are susceptible to biological colonization. The dynamics of the biological component can generate problems for the conservation of architectural components (Fig. 59.1).

1750

R. Manganelli Del Fa` et al.

Fig. 59.1 Examples of (a) incompatibility between vegetal and constructed elements; (b) biological colonization; and (c) growth of plants on the surface. (Photo ISPC-CNR Archive)

The open environment with its characteristic microclimatic conditions, besides the intrinsic bio-receptivity of the stone materials, are key factors that influence the settlement and development of micro- and macroorganisms. The presence of the profuse vegetation induces peculiar effects at a microenvironment level such as [9]: – Decrease of the temperature values and increase of the relative humidity values for the areas located outside of the garden (a few Celsius degrees less and about 10% more of RH), and there is an increase in areas without direct sunlight. The wind is reduced as well, contributing to keeping colder and wetter conditions. – A reduction in chemical pollution due to the presence of tree foliage, with positive effects on the conservation of the stone materials, but favoring the development of the organisms which are not pollution sensitive, such as lichens, many of them being air pollution indicators. These peculiar microclimatic conditions favor microbiological development and the microorganism’s biodiversity is higher than in other open-air environments. The presence of different fountains or decorative water sources represents another type of micro-ecosystem, which is usually placed in sun-illuminated spaces [10]. The conservation problems of these artifacts are very complex, having parts completely immersed in water and others completely dry, with different exposition.

59

Sustainable Conservation and Restoration of Historical Gardens

1751

These different situations induce a different distribution of the organisms on the artifacts according to their specific ecological requirements. The fountains are more often subjected to restoration interventions than the other decorative artifacts present in the historical gardens. The biodiversity present on the stone materials decorating the historical gardens gives rise to the formation of alteration phenomena such as biofilms, spots, patinas, incrustations, pitting, etc. Green and green-reddish patinas or biofilm are formed in the wettest surfaces, due to the development of phototrophic microorganisms such as algae, cyanobacteria, and diatoms. The black patinas are mainly due to the presence of cyanobacteria and fungi, while the reddish ones are formed by some algal colonization containing a high quantity of carotenoid pigments. Plenty of lichens, epilithic, and/or endolithic, as well as different mosses are frequent on the stone surfaces and often form large coverings. The presence of guano can be observed on horizontal surfaces. Besides this organic material source, the natural degradation and organisms succession (such as leaves, dead microorganisms, etc.) favor the eutrophication phenomena of the stone and contribute to its deterioration. The stone materials used for garden architecture and decorations are usually local and not very expensive ones. The lithotypes with a rough surface and higher water retention, such as tuffs, sandstones, or marbles (carbonatic stone in general) are easily colonized by the microorganisms due to the soiling deposition phenomena, enriching the surface with soluble minerals that can be further used as a nutrient or incrustation formations that increase the porosity of the surface, creating, therefore, and favorable living conditions. Often, the stone artifacts are subject to restoration interventions; hence, the natural ability to be colonized is changed, this type of bio-receptivity being called tertiary bio-receptivity [11]. This kind of secondary colonization is influenced by the nearby presence of eurytopic species, which easily adapt themselves to the restored surfaces. However, it has been said that the presence of mosses and lichens on the garden stone materials was very appreciated for a long time and even at the current time, some people give an ornamental and aesthetic value to the biological colonization. In the past, even the damage produced by the vegetation’s roots was little considered, to the detriment of the stone material deterioration. Nowadays, the conservation of flora biodiversity is preferred to the stone material safety only in special circumstances, when, for example, there is some biocenosis with special naturalistic value, such is the case of the Eucladio-Adiantetum biocenosis from the Nymphaeum Pioggia al Palatino from Rome [10].

59.3

Stone Furniture and Decay Processes

Weathering and deterioration of stones are a natural occurrence. Since deterioration is a complex process, many words are used to describe it. For example, weathering is used for any chemical or mechanical process by which stones exposed to the weather undergo changes in character and deteriorate, while deterioration implies the impairment of value and use. On the other hand, alteration is defined as a

1752

R. Manganelli Del Fa` et al.

modification of the material that does not necessarily imply a worsening of its characteristics for conservation. Degradation and decay, instead, imply a change for the worse (in terms of modification of the intrinsic chemical and physical stone properties) and tend to be used interchangeably. In studies on stone deterioration and conservation, terminological confusions lead to major communication problems between scientists, conservators, and practitioners. In this context, it is of primary importance to set up a common language, since many terminologies can be a source of misunderstanding. In 2008, the ICOMOS (International Council on Monuments and Sites) International Scientific Committee for Stone (ICSC) constituted an illustrated glossary on stone deterioration that represents an important tool for scientific discussions on decay phenomena and processes [12]. The understanding of weathering processes and of the contribution of various intrinsic (e.g., microfabric, texture, chemical, and mineralogical composition) and extrinsic (e.g., temperature, moisture, freeze-thaw cycles, salt crystallization, biological colonization) factors to it is a crucial precondition for both the choice of appropriate conservation and restoration strategies of deteriorated stone artifacts. Among the main stone materials employed in the historical gardens architecture, marbles and sandstones are the most used.

59.3.1 Marble Marble is a non-foliated metamorphic rock that forms when limestone (or dolomite rock) is subject to the heat and pressure of metamorphism. Metamorphism causes variable recrystallization of the original carbonate mineral grains. The resulting marble rock is typically composed of an interlocking mosaic of carbonate crystals. Common white marble, such as Apuan one coming from Carrara quarries, can present several typologies; from white (statuary type) to more or less veined in compositional terms, it consists of 99% calcite and traces of quartz, albite, muscovite, and pyrite. Common marble can also show microstructures varying from polygonal granoblastic with rectilinear contacts between grains, to xenoblastic with saturated contacts between grains [13, 14]. Numerous studies have been performed to identify the principal weathering factors for the marbles. It has been demonstrated that temperature changes (thermoclastism) play an important role in marble alteration, leading to decohesion among the crystals. Experimental observations have shown that a temperature increase of only 20–30  C is sufficient to produce partial decohesion of calcite grains [15]. This is favored by the strong anisotropy of the thermal expansion coefficient of calcite [16–18] and is conditioned by the type of microstructure [19, 20]. Calcite is the only mineral that upon heating expands in one direction while contracting in the other; and upon cooling, it will contract along the c axis while expanding along with the other ones. Therefore, the stress induced by heating leads to fissuring and, eventually, fracturing and results in an increase in porosity [21]. This may already occur at temperatures around 40–50  C, a value that is easily reached by a stone

59

Sustainable Conservation and Restoration of Historical Gardens

1753

surface on summer days, even in northern countries. But then, the cooling cycles that occur in winter will also contribute to grain decohesion. This phenomenon has also been observed in marble quarries, and the deteriorated marble is referred to as marmo cotto [16]. Moreover, other types of decay phenomena are due to chemical processes such as the dissolution of calcite by meteoric water and its associated acidity [22] and sulphation, with consequent possible formation of well-known black crusts occurring on surfaces not directly exposed to rainfall [23, 24]. These crusts, in addition to causing a visually unattractive chromatic alteration, favor decay, as demonstrated by the extremely deteriorated conditions of the marble underlying the crusts.

59.3.2 Sandstone Sandstones are by far the most common clastic sedimentary rocks, composed mainly of sand-sized (0.0625 to 2 mm) mineral particles or rock fragments that are used in architecture and objects expressing the cultural heritage. Sandstones are formed in a wide range of environments from water (as in a stream, lake, or sea) or air (as in a desert). Typically, sedimentation occurs by the sand settling out from suspension; finally, once it has accumulated, the sand becomes sandstone when it is compacted by the pressure of overlying deposits and cemented by the precipitation of minerals within the pore spaces between sand grains. The most durable sandstones are the silica cemented quartz sandstones that often contain feldspars such as orthoclase or plagioclase. The presence of clay minerals, instead, can negatively influence the properties and durability of sandstones. Especially swelling clays, such as montmorillonite (smectite group), can harm the use and long-term behavior of sandstone. Some sandstones are primarily composed of carbonate grains that in most cases are less durable than quartz sandstones, although the workability is much better than their siliceous ones. Decay phenomena of this stone are essentially linked to natural physical and physicochemical processes due to the concomitant action of two factors: sudden temperature changes and humidity [25–27]. Water can act in several ways, by washing away the clay matrix via a purely mechanical washing action, making the stone completely disaggregated, by swelling the lattice of the clay minerals leading to surface exfoliation and disintegration [28]. Moreover, the water can cause dissolution and reprecipitation of the calcitic cement, giving rise to the formation of cohesive crusts with low porosity that, being discontinuous with the substrate, tend to fall and to be reformed with the progressive destruction of the architectural element. Also, the freezethaw cycles of water leading to the disintegration of the stone must be considered.

59.4

Evaluation of Decay: Multidisciplinary Approach

The analytical evaluation of the extent of the deterioration of the material, similar to the definition of the chemical-physics processes that regulate them, is a key parameter for the definition of any conservation protocol, optimizing the efforts to avoid expensive restoration interventions in emergency conditions.

1754

R. Manganelli Del Fa` et al.

The use of a multi-analytical and multidisciplinary approach allows us to address various aspects: knowledge of the materials, evaluation of the state of conservation of the materials, identification of the most suitable conservation practices, and monitoring of the stone surfaces and environmental parameters. In terms of verifying the state of conservation of the materials, it is known that all natural materials exposed to conditions different from the original ones tend to reach a new equilibrium, through chemical-physical transformations in nature. These transformations can lead to significant alterations and deteriorations of the material. To identify the deterioration mechanisms, the intrinsic characteristics of the material have to be taken into account and the external factors that trigger and determine the alteration and the decay have to be assessed.

59.4.1 Laboratory Analytical Techniques The sampling of fragments of the stone artifacts for laboratory analyses is a critical step as it should be well representative and less invasive as possible. Among several analytical methodologies [29, 30], X-ray diffractometry (XRD) identifies the mineralogical composition (only crystalline components) inside the sample itself. This technique can be used to assess the state of the alteration and deterioration phenomena (identification of salt, precipitation phases/alteration crusts) as well as to perform a characterization of the materials (e.g., stones, plasters, mortars) [31]. Only a small amount of sample (few mgs) is required. Moreover, because XRD is a nondestructive technique, the sample is not subject to any modification during the analysis and then can be used for other analytical methodological applications. Fourier transform infrared spectroscopy (FT-IR), associated with the XRD, led to the identification both of crystalline and noncrystalline organic matters (e.g., oils, resins, glues) and numerous inorganic compounds (e.g., carbonates, sulfates, silicates, oxalates, some pigments, etc.). This technique is based on the properties of functional groups (carbonates, sulfates, etc.) to absorb infrared radiation (IR) of specific wavelengths [32]. Moreover, certain samples, conveniently selected, can be embedded in consolidating resin to obtain thin and cross sections that can be analyzed by microscopic and electronic techniques. Examining the samples under the microscope not only enabled magnified images of the sample to be seen but also gave compositional information and the current state of conservation (e.g., the degree of detachment among the grains in a marble stone). A transmitted light microscope used to examine thin sections (about 30 μm) of stones (but also mortars and plasters) led to the analysis of the principal compositional, textural, and microfabric parameters. In terms of the stone materials, using such a technique, it is possible to define the mineralogical-petrographic classification and the state of conservation and to provide general information on the provenance [18]. In the case of mortars, by examining these thin sections, it is possible to obtain information on the raw materials used for their preparation (such as the limestone used for the binder),

59

Sustainable Conservation and Restoration of Historical Gardens

1755

the type and provenance of the sand used as aggregate, and information of technological nature, which is useful in recreating the original manufacturing techniques [33]. A reflected light microscope, instead, allows the observation of cross sections that is very useful for the description of the morphological characteristics of the sample and the evaluation of the stratigraphic sequence (e.g., painted layers, types of mortar, superficial deposits and crusts, protective treatments) and of the state of conservation. The use of some types of light (i.e., UV light) enables us to identify the presence of organic material in the single layer. Cross sections, and also thin sections, can be examined by scanning electron microscope equipped with elemental analysis (SEM-EDS) to gather morphological and compositional information [34, 35] (Fig. 59.2).

59.4.2 Biodeteriogen Investigation To investigate the presence of the biodeteriogens, in particular for checking the metabolic activity of biodeteriogens infesting the surface of stone, two useful methods can be used. The first method is based on the quantification of the adenosine triphosphate (ATP) present on a certain surface [36]. The ATP is a molecule involved in the energetic processes of the cells and is present only in the alive cells, metabolically active. The instrument that measures the quantity of ATP in a certain sample is called a luminometer or bioluminometer. The second is based on the evaluation of the photosynthetic system efficiency in phototrophic organisms growing on the stone surface. In the last decades, the pulse amplitude modulated fluorescence (PAM) technique has been adopted for in situ investigation, recently a version of PAM for image acquisition can provide information about the photochemistry of the chlorophyll-containing species on portion of surface up to several tens of square centimeters [37]. It gives information both about the evolution of phototrophic colonization in time and about the effect of control treatments on biodeteriogens as regards their ability to produce photosynthesis, hence as regards their vitality condition. Regarding the study of the microbe-substrata mineral interaction, microinvasive sampling needs to be performed, for the morphological characterization of the biological organisms present on stone materials (Fig. 59.3). The little stone fragments containing organisms can be analyzed with different microscopic techniques (epifluorescence microscopy, reflectance microscopy, environmental scanning electron microscopy). The samples can be observed as such or may be included in an epoxy-resin which gives the possibility to study a cross-section of the sample, and on which some specific stains can be used to reveal the fungal hyphae inside the stone. The epifluorescence microscopy by using UV filters is very useful for observing the autofluorescence of the chlorophyll-containing microorganisms, such as green algae and cyanobacteria. These techniques are of great help for analyzing the deepness of the biological colonization (phototrophs and heterotrophs) within the stone, bringing essential information for defining a conservation treatment.

R. Manganelli Del Fa` et al.

1756

Fig. 59.2 Examples of (a) mechanical cleaning of biological patina by brush and (b) microsampling of a stone fragment; stereomicroscopic view of marble fragments affected by (c) lichens (Verrucaria sp.) and (d) yellow lichens and black patina made of cyanobacteria; microphotographs of thin sections observed in transmitted light of an (e) Greek marble (dolomitic, medium-coarse grain size, heteroblastic microstructure, curved grain boundary shapes), and an (f) Carrara marble (calcitic, very fine grain size, homeoblastic microstructure, straight grain boundary shapes), both in crossed nicols. (Photo ISPC-CNR Archive)

59.5

Sustainability and Minimal Intervention

The term and concept of sustainability have come to be widely used in cultural heritage studies, and their usefulness in conservation has also been recognized (cit.) [38].

59

Sustainable Conservation and Restoration of Historical Gardens

1757

Fig. 59.3 Phototrophic biodeteriogens on a sandstone fragment observed in (a) reflected light and (b) epifluorescence mode (the red color indicates the presence of the chlorophyll); (c) microscopic view of different species of microorganisms present on a marble fragment (1: Alternaria sp. – a fungus; 2: Gloeocapsa sp. – a cyanobacterium; 3: Chlorophyceae – a green alga); (d) microscopic view of a lichen, often observed on stone materials exposed in historical gardens. (Photo ISPC-CNR Archive)

Sustainability [39] assumes many meanings in the field of restoration, mostly related to energy aspects (e.g., restoration of buildings and their re-use), and economic and ecological ones (e.g., for restoration products). Conservation is mainly focused on the concept of eco-sustainable conservation, studying interventions and products with a low or null impact on the environment and restorers. For the economic aspect of conservation, the concept of sustainable development has been developed. A conservation project should help the local economy by generating growth and income. In the end, it is possible to speak about social sustainability when conservation enhances the quality of life and local communities and society in general [40]. Considering the range of meanings that sustainability has acquired in conservation, it is clear that finding a univocal definition of the concept is a challenging task. Throsby [41] and Avrami [42], in their works, defined sustainability as a set of benchmarks or an organizational framework. According to these definitions, it would be more useful to do choices that end up in sustainable results, rather than trying to find a unique definition of sustainability.

1758

R. Manganelli Del Fa` et al.

With this in mind, it is possible to consider sustainability as a set of principles, a collection of guidelines to follow in the decision-making process in conservation. Therefore, sustainable principles comprehend a wide range of aspects; the aim of sustainability is achieved by paying attention to environmental, economic, and social factors. Furthermore, these principles should be integrated into the principles of conservation and should constitute the final goal of the whole process. How is possible to integrate the sustainable concept into a conservation project? It is possible with a comprehensive planning project, a project that plans all the actions to take up with a long-term perspective. The Codice dei Beni Culturali e del Paesaggio (Cultural Heritage and Landscape Code, Legislative Decree 42/2004), paragraph 29, recite the preservation of the Cultural Heritage is ensured through a coherent, coordinated, and planned activity of study, prevention, maintenance, and restoration. Moreover, the Decree defines Restoration as “the complex of operations on the work of art, aimed at maintaining its material integrity and recovering it, protecting and transmitting its cultural values.” Furthermore, paragraph number 29 claims that in the case of risks for the work of art (e.g., loss of materials or parts), the restoration will also include structural improvement works and make it safe. The requirement of minimal intervention with a low impact on the artworks and for the environment is the first target towards sustainable conservation. It is important to minimize the restoring action in order to reduce the stress to the work of art during each intervention and preserve all the information about the original installation and history of an artifact. Historic garden conservation requires various skills to manage green and architecture that interact with each other in a very complex system. The garden must be considered a unique resource, where the two components (green and architecture) are synergistically integrated but require different kinds of protection and maintenance. In a restoration project, analyzing the historical data relating to the work of art (history and conservative events) is necessary. Furthermore, a thorough knowledge of the constituent materials and their state of preservation is required. In the end, the technical aspects linked to the installation of the material must not be overlooked. The historical information research must be carried out in archives, libraries, public and private offices, and also through oral testimonies; constituent materials will be investigated both with the direct observation and with diagnostic investigations; the executive techniques will be evaluated both with the traces of tools used for its realization and with its intrinsic characteristics (dimensions, weight, positioning); the state of conservation of the work will be established by direct observation of the effects produced by the various degradation phenomena, both superficial and structural; for the restoration, scientific investigations carried out in the past can be compared with newly acquired data. The proposed methodology suggests minimal intervention, called shy restoration (restauro timido). The architect Marco Ermentini theorized the shy restoration in his writings on restoration and architecture, declaring this concept in his book Manifesto rosso dell’architettura timida [43].

59

Sustainable Conservation and Restoration of Historical Gardens

1759

All phases of a restoration project are indispensable and sometimes urgent, for example, when the stability of an object is damaged. The Manifesto Rosso book analyzes the cleaning phases of the restoration, capable of characterizing the final result. Keeping in mind the conservation requirement, it is advisable to ask if the complete removal of the biological patina from the stone furnishings of the gardens is necessary or not. Is it right to completely change the appearance of stone furniture, eliminating the patina developed over time on the surface, as it was probably planned? One of the main phenomena that can be found on the garden’s artifacts is the presence of microflora. The removal of this kind of patina, according to the minimal intervention, is that of a gradual removal to be carried out without the use of biocidal products but with simple tools and techniques like small wooden sticks, synthetic and natural brushes, slightly abrasive sponges, scalpel, demineralized water, and without its total elimination which would lead to an excessive visual contrast with the environmental context. The large wealth of the shy restoration, and in general the shy architecture, is the absence, the renunciation of the operation according to the principle of peace stands still, and the uselessness of operation if not extremely necessary. Its trait is to hide, stopping itself at the opportune moment, not the spectacle of the operation, the knowledge of not understanding it all, and the prudence, in other words, the shyness (cit.) [43].

59.6

Conclusion

The main characteristic of a garden, historical or not, is its intrinsic trend to change over time, so the aspect of the garden is a consequence of the interaction among plant growth, stone artifacts decay, and environmental and anthropic factors (visitors, pollution, etc.). The conservation status of trees can become precarious, due to both their age and the effect of a rapidly changing climate. Moreover, the natural processes of senescence, various types of biotic and abiotic stresses, and extreme events associated with global warming (e.g., droughts and wind storms) have recently caused serious injuries to many old trees in the last decade. As a consequence, the physiological conditions and stability of many trees put stone artifacts at risk of (accidental) damage. The natural decay processes of stone are due to environmental factors (acid rain, wind, freeze-thaw cycle, particulate, etc.) that affect the surface and activate physicochemical degradation phenomena. All the activities that are focused on slowing down these processes are classified as restoration work. In this contribution, the author’s aim was the application of particular practices and procedures to reduce or to mitigate the impact of the restoration work on the environment. The impact on the environment must be as little as possible in terms of aesthetical and historical meaning in the perspective to maintain the original parts and patina and, at the same time, must balance the equilibrium among plants (trees,

1760

R. Manganelli Del Fa` et al.

vegetation, and lichens or algae on stone surfaces), stone constructed elements (sculptures, fountains), and stone furniture (benches). In this contribution, the main operational phases for the restoration are examined concerning the criteria presented in Marco Ermentini’s Manifestino rosso dell’architettura timida. The cleaning is certainly the basic step because it characterizes the final result of the restoration of stone furniture in a historical garden. The importance of conceptualizing works of art according to their decorative function, subsequent to the architectural context for which they were created or later adapted, is necessary to define a homogeneous intervention. A condition to talk about conservative restoration of stone furnishing is that the alterations on the surface of their facies is left unaltered. The biodegradation is certainly the main decay phenomenon on stone artifacts. The minimal intervention is our proposal of approach to restoration that through a gradual removal, without the use of biocidal products, seeks to avoid a total cleaning, which would lead to an excessive visual contrast with the environment. Its target is to stop at the right moment, the non-spectacularization of the restoration, and the prudence, i.e., the shyness towards the artifacts and the environment where they are found, keeping in mind that restorers are unable to understand and control every aspect or phenomenon. The balancing of the works with their environment is possible following the procedures and best practices suggested by the shy restoration using treatments that preserve the artifacts from rain – and ensuing physical-chemical events, such as loss of cohesion and excessive proliferation of biological patina. The effects of these treatments are not intended as once and for all, but they must be periodically repeated over time following an established maintenance program based on monitoring the effectiveness and durability of the intervention because gardens are a witness of passing time that has to be maintained for future generations.

References 1. Moggi G (2005) Il giardino rinascimentale. In: Caneva G (ed) La biologia vegetale per i beni culturali, vol 2. Nardini Editore, Firenze 2. Accati E, Devecchi M (2005) Il restauro del giardino storico (Review). Italus Hortus 12(4):17–30 3. Comitato internazionale dei giardini storici ICOMOS-IFLA: carta dei giardini storici detta “carta di Firenze”. Proc. Il Giardino storico. protezione e restauro, Atti del VI Colloquio internazionale sulla protezione ed il restauro dei giardini storici MAY 1981. Regione Toscana ICOMOS, Firenze 1987, pp 201–206 4. Belli Barsali I (1987) Per una carta del restauro dei giardini storici. In: Proc. Il Giardino storico. protezione e restauro, Atti del VI Colloquio internazionale sulla protezione ed il restauro dei giardini storici MAY 1981. Regione Toscana ICOMOS, Firenze, pp 159–166 5. Galletti G, Matteini T (2005) Tra Rinascimento e manierismo: il giardino del Cinquecento. In: Giardini. L’arte del verde attraverso i secoli. Giunti Editore, Milano, pp 36–71 6. Maniglio Calcagno A (2006) Architettura del paesaggio. Evoluzione storica. Franco Angeli editore, Milano 7. Zoppi M (2009) Storia del giardino europeo. Alinea Editrice, Firenze

59

Sustainable Conservation and Restoration of Historical Gardens

1761

8. Galletti G, Matteini T (2005) Dal giardino barocco al parco paesaggistico. In: Giardini. L’arte del verde attraverso i secoli. Giunti Editore, Milano, pp 72–109 9. Caneva G, Piervittor R, Roccardi A, Tomaselli ML (2007) Monumenti e manufatti in parchi e ambienti rurali. In: Caneva G, Nugari MP, Salvadori O (eds) La biologia vegetale per i beni culturali, vol 1. Nardini Editore, Firenze, pp 193–196 10. Pietrini AM, Ricci S (2007) Fontane e Ninfei. In: Caneva G, Nugari MP, Salvadori O (eds) La biologia vegetale per i beni culturali, vol 1. Nardini Editore, Firenze, pp 199–205 11. Guillitte O (1995) Bioreceptivity: a new concept for building ecology studies. Sci Total Environ 167(1–3):215–220. https://doi.org/10.1016/0048-9697(95)04582-L 12. Vergès-Belmin V (eds) (2008) Illustrated glossary on stone deterioration patterns, Monuments and Sites XV. ICOMOS Documentation Centre, Paris 13. Molli G, Heilbronner Panozzo R, Wittensoldener V (1997) Microstructural features of naturally deformed Alpi Apuane marbles: an introduction. Proceedings of conference on Deformation Mechanism in Nature and Experiments, Basel, Switzerland, 17–19 March 1997. Elsevier, Amsterdam, pp 45–46 14. Cantisani E, Fratini F, Malesani P, Molli G (2005) Mineralogical and petrophysical characterisation of white Apuan marble. Periodico di Mineralogia 74(2):117–140 15. Battaglia S, Franzini M, Mango F (1993) High sensitivity apparatus for measuring linear thermal expansion: preliminary results on the response of marbles to thermal cycles., Il. Nuovo Cimento C 16(4):453–461 16. Bertagnini A, Franzini M, Gratziu C, Spampinato M (1984) Il marmo cotto in natura e nei monumenti. RENDICONTI Società Italiana di Mineralogia e Petrografia 39(1):39–46 17. Koch A, Siegesmund S (2004) The combined effect of moisture and temperature on the anomalous expansion behavior of marbles. Environ Geol 46(3):350–363 18. Cantisani E, Pecchioni E, Fratini F, Garzonio CA, Malesani P, Molli G (2009) Thermal stress in the Apuan marbles: relationships between microstructure and petrophysical characteristics. Int J Rock Mech Min Sci 46(1):128–137 19. Barsottelli M, Fratini F, Giorgetti G, Manganelli Del Fá C, Molli G (1998) Microfabric and alteration in Carrara marble: a preliminary study. Sci Tech Cult Heritage 7(2):115–126 20. Cantisani E, Canova R, Fratini F, Manganelli Del Fá C, Mazzuoli R, Molli G (2000) Relationships between microstructures and physical properties of white Apuan marbles: inferences on weathering durability. Periodico di Mineralogia 69(3):257–268 21. Malaga-Starzec K, Lindquist JE, Bjorn S (2002) Experimental study on the variation in porosity of marble as function of temperature. In: Siegesmund S, Weiss T, Vollbrecht A (eds) Natural stone, weathering phenomena, conservation strategies and case studies. The Geological Society, London, pp 81–88 22. Reddy MM (1988) Acid rain damage to carbonate stone: a quantitative assessment based on the aqueous geochemistry of rainfall runoff from stone. Earth Surf Process Landf 13(4): 335–354 23. Camuffo D, Del Monte M, Sabbion C, Vittori O (1982) Wetting, deterioration and visual features of stone surfaces in an urban area. Atmos Environ 16(9):2253–2259 24. Charola AE, Puhringer J, Steiger M (2007) Gypsum: a review of its role in the deterioration of building materials. Environ Geol 52(2):339–352 25. Tukington AV, Smith BJ, Basheer PAM (2002) The effect of block retreat on subsurface temperature and moisture conditions in sandstone. In: Prikryl R, Viles H (eds) Proc. understanding and managing stone decay. The Karolinum Press, Prague, pp 113–126 26. Torok A, Hajpal M (2005) Effect of temperature changes on the mineralogy and physical properties of sandstones, A laboratory study. Restorat Build Monuments 11(4):211–218 27. Fratini F, Rescic S (2014) The stone materials of the historical architecture of Tuscany, Italy. Special Publications 391(1). Geological Society, London, pp 71–92 28. Anderson RL, Ratcliffe I, Greenwell HC, Williams PA, Cliffe S, Coveney PV (2010) Clay swelling – a challenge in the oilfield. Earth Sci Rev 98(3–4):201–216

1762

R. Manganelli Del Fa` et al.

29. Artioli G (2010) Scientific methods and cultural heritage: an introduction to the application of materials science to archaeometry and conservation science. Oxford University Press, New York 30. Varella EA (ed) (2013) Conservation science for the cultural heritage: applications of instrumental analysis, Lecture notes in chemistry, vol 79. Springer, Berlin 31. Corbeil MC (2004) Applications of x-ray diffraction in conservation science and archaeometry. Adv X-ray Anal 47(2):18–29 32. Derrick MR, Stulik D, Landry JM (1999) Infrared spectroscopy in conservation science. Getty Conservation Institute, Los Angeles 33. Pecchioni E, Fratini F, Cantisani E (2014) Atlas of ancient mortars in thin section under the microscope. Nardini Editore, Firenze 34. Doehne E, Stulik D (1990) Application of the environmental scanning electron microscope to conservation science. Scanning Microsc 4(2):275–286 35. Lanterna G (1999) L’uso del SEM nella scienza della conservazione/the use of SEM in science for conservation. OPD Restauro 11:39–58 36. Ranalli G, Zanardini E, Pasini P, Roda A (2003) Rapid biodeteriogen and biocide diagnosis on artwork: a bioluminescent low-light imaging technique. Ann Microbiol 53:1–13 37. Osticioli I, Mascalchi M, Pinna D, Siano S (2013) Potential of chlorophyll fluorescence imaging for assessing bio-viability changes of biodeteriogen growths on stone monuments. In: Proceedings of the SPIE 8790, Optics for Arts, Architecture, and Archaeology IV, 879003. https:// doi.org/10.1117/12.2020563 38. Muñoz-Viñas S (2012) Contemporary theory of conservation. Stud Conserv 47(1):1–239 39. Pretelli M, Signorelli L (2014) Sostenibilità tra economia e cultura. Restauro, ri-costruzione e “riuso” in alcuni casi del secondo dopoguerra tedesco (tra est e ovest), Proc. Scienza e beni Culturali XXX Quale sostenibilità per il restauro? Edizione Arcadia Ricerche, Bressanone, pp 41–51 40. Guiducci F (2014) A sustainable approach to site conservation: conservation of excavated structures in Dangeil, Sudan. Proceedings of the Scienza e beni Culturali XXX Quale sostenibilità per il restauro? Edizione Arcadia Ricerche, Bressanone, pp 201–212 41. Throsby D (2003) Sustainability in the conservation of the built environment: an economist’s perspective. In: Teutonico JM, Matero F (eds) Proceedings managing change: sustainable approaches to the conservation of the built environment, 4th Annual US/ICOMOS INternational symposium. Getty Publications, Los Angeles, pp 3–10 42. Avrami E (2012) A system approach to historic preservation in an era of sustainability planning, doctoral dissertation. The state university of New Jersey 43. Ermentini M (2008) Restauro timido. Architettura, affetto, gioco. Nardini, Firenze

Geological Structural Analysis Applied to Archaeoseismology

60

Jorge Luis Giner-Robles, Miguel A´ngel Rodríguez-Pascua, Rau´l Pe´rez-Lo´pez, Pablo Gabriel Silva, Teresa Bardají, Elvira Roquero, Javier Elez, and María A´ngeles Perucha

Contents 60.1 60.2 60.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Methodology and Rationale . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Building Oriented Damage (BOD): Archaeological Seismic Structural Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60.3.1 Damage Identification and Inventory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60.3.2 Strain Seismic Damage Quantification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1764 1765 1767 1768 1768 1776 1776

Abstract

The spatial and geostatistical analysis of the building damage caused by earthquakes in archaeological sites provides significant information about ancient and

J. L. Giner-Robles (*) Dpto. Geología y Geoquímica. Fac. Ciencias, Universidad Autónoma de Madrid (UAM), Madrid, Spain e-mail: [email protected] M. Á. Rodríguez-Pascua · R. Pérez-López · M. Á. Perucha Instituto Geológico y Minero de España (IGME), Madrid, Spain e-mail: [email protected]; [email protected]; [email protected] P. G. Silva · J. Elez Dpto. Geología, Universidad de Salamanca, Avila, Spain e-mail: [email protected]; [email protected] T. Bardají U.D. Geología, Universidad de Alcalá (UAH), Madrid, Spain e-mail: [email protected] E. Roquero Dpto. de Edafología, E.T.S.I. Agrónomos, Universidad Politécnica (UPM), Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_60

1763

1764

J. L. Giner-Robles et al.

historic seismic events such as earthquake directivity, seismic maximum intensity, and the spatial effect of seismic surface wave propagation. A comprehensive mapping of the building damage provides seismic intensity data that can be used in further hazard studies aimed at the preventive protection of such cultural heritage. This chapter describes a method to quantify building oriented damage recorded in archaeological sites to establish the spatial seismic pattern of damage, leading to the further identification of the potential seismic source location. The inventory and mapping of the earthquake archaeological effects depict the preferred arrangement of building oriented damage indicating the most probable direction of seismic ground shaking. These data can be summarized in rose diagrams of ground shaking for individual buildings, one archaeological site, or a set of archaeological sites affected by an ancient earthquake. The quantification of the ground seismic deformation of the sites follows similar procedures to that used in structural geology to determine the tectonic maximum horizontal shortening relative to active tectonic fields, but in this case, it focused to identify the seismic origin and the potential location of the causative seismic source. This chapter describes the application of this methodology to ancient and historical earthquakes in South Spain (e.g., Roman archaeological site of Baelo Claudia; Cádiz) but also to the destructive instrumental earthquake of Lorca, 2011 (5.2 Mw), recently occurred in SE Spain (Murcia). The application of this methodology to ancient, historical, and instrumental earthquakes ( VII EMS-98) allows the cross-checking and validation of the proposed techniques for the analysis of past earthquakes witnessed in the cultural heritage.

60.1

Introduction

The completeness of the seismic catalog is one of the main challenges in the analysis of seismic hazard in areas that undergo a low to medium rate of seismic activity (i.e., one earthquake M 6 sized every 100 years). The incorporation of ancient or historical earthquake data in these catalogs is a major improvement in the knowledge of seismic cycle (time between characteristic earthquake occurrence), by the enlargement of the time period of the recorded seismic events within a tectonic zone. However, available descriptions of the environmental effects and the building damage produced by historical seismic events are usually poorly described making it necessary complementary analyses to obtain quantitative seismic data for further earthquake hazard analysis. There is a well-known description of building damage affecting popular archaeological and holy sites in the Middle East interpreted by different authors as seismic following several biblical passages [1–3]. Also, there are methodological contributions that explain the seismic effects of past and destructive events in historical buildings or ancient remains, carried out by Nikonov [4], Stiros [5], Marco [6] and Ambrasseys [7]. A key document entitled “Archaeoseismology” [8] was the pioneer in this new discipline, introducing archaeoseismology as the study of historic earthquakes affecting ancient cultures and past civilizations, even as breakpoints

60

Geological Structural Analysis Applied to Archaeoseismology

1765

responsible for cultural decadence and eventual societal collapses. Silva et al. [9] are the first authors using structural geology techniques (i.e., rose diagrams) to quantify building damage (fractures, cracks, displacements, etc.) in archaeological sites. Later, Sintubin et al. [10] print the label of “ancient earthquakes” to identify those seismic events out of the historical writing records, only witnessed in the archaeological remains susceptible to be quantified. The relevance of this emergent discipline is the introduction of a new scientific approach, combining archaeology, geology, and geophysics, in the search of ancient earthquakes. This provides scientific data for the occurrence of ancient destructive earthquakes, normally previously unknown, allowing the enlargement of conventional seismic catalogs and the improvement of future seismic hazard analyses. This is of special interest in zones with apparent low seismicity in which the historical writing records are not specially detailed, and archaeoseismological data help to put constraints to seismic damage. This is the case of the Western Mediterranean region during ancient times. Most of the published papers point to the occurrence of specific spatial patterns of destruction (e.g., orientation of fallen walls and columns) as one of the key criteria to identify seismic damage in archaeological sites. This is the identification of building oriented damage (BOD) in archaeological remains, which helps to identify the archaeological pattern of destruction, helping for their future preservation. Several authors [11–15] analyzed BOD on buildings, monuments, and other man-made structures linked to instrumental earthquakes. These studies are relevant since they allow the cross-checking of BOD of historical events with instrumental earthquakes. Instrumental records provide well-constrained source parameters (i.e., magnitude, depth, focal mechanism solutions, etc.) to cross-check patterns of oriented damage with those proposed for archaeological sites. Hence, parameters such as intensity and the seismic source location can be estimated for historical earthquakes from archaeoseismic analysis. This work describes a new methodology based on standard structural geology analyses for mapping the spatial deformation and kinematics related to a tectonic stress field (i.e., strain analysis) [16, 17], here simplified to the stress release of a single earthquake. The measurement of the set of seismic deformations defined by earthquake archaeological effects (EAEs) [14] can be summarized in “rose diagrams” identifying single vector (1D orientation; e.g., fallen columns) or planar data (2D orientation; e.g., tilted walls). This analysis obtains and represents BOD featuring the preferent orientation of the ground movement. In recent earthquakes, this analysis must be performed a few days after the earthquake, because of the subsequent works for reconstruction of the affected zones.

60.2

Methodology and Rationale

The methodology of the seismic analysis for BOD consists of the strain quantification analysis (Fig. 60.1), as the first step for obtaining oriented data and statistical information about the frequency, number, and spatial distribution of the seismic damage. The type of data used (Fig. 60.1) involves the analysis and parameterization

1766

J. L. Giner-Robles et al.

Fig. 60.1 Outline of the methodology proposed for the incorporation of historical and noninstrumental earthquakes in seismic catalogs. This methodology combines different types of data (EEEs and EAEs) for analyzing, characterizing, and parameterizing earthquakes. The recognition of these data allows determining the intensity of environmental, geological, and building effects of earthquakes. (Modified from Giner-Robles et al. [31])

of the geological environmental earthquake effects (EEE) [18] and the earthquake archaeological effects (EAE) [14]. These data are based on the work of previous authors concerned in the searching of common effects of earthquake intensity and geological effects: (1) Michetti et al. [18] classify and parameterize the different geological effects of earthquakes (EEE), defining a macroseismic intensity scale, environmental seismic intensity (ESI-07). The advantage is the possibility to apply this scale in paleo-, historical, and instrumental earthquakes; (2) Rodriguez-Pascua et al. [14] establish a classification of the different archaeoseismological effects produced by earthquakes (EAEs), describing how the building and the ground respond to the ground shaking: tilted walls, collapsed vaults in cathedrals and churches, oriented fallen columns, and collapse of arches and lintels. These authors also mapped all of these effects to figure out and discriminate other possibilities for the destruction horizon; (3) Giner-Robles et al. [19] propose an “ad hoc” protocol to discriminate the seismic origin of the deformations found in archaeological sites from other effects such as war or collapse for abandonment, among others; (4) Rodríguez-Pascua et al. [20] propose a seismic intensity scale based on EAEs

60

Geological Structural Analysis Applied to Archaeoseismology

1767

and divided into 12 numbers, related to the type of construction and type of damage; (5) Silva et al. [21], finally, make a correlation between seismic intensities and peak ground accelerations for ancient and historical seismic events. The modeling of these data is based on the ShakeMaps of the USGS Earthquake Hazards Program [22], and it allows us to define the seismic source (orientation, type of movement, and surface earthquake rupture) and the earthquake size. The power of this technique is the possibility to estimate these data from historical earthquakes where only intensity data from population affected are available [23]. Giner-Robles et al. [19] introduced a detailed description of each of the EAEs types from the seismic damage recognized in the Roman city of Baelo Claudia (Cadiz, Spain) [9]. In Baelo Claudia different authors [9, 24–26] defined the occurrence of two destructive earthquakes (AD 40–60 and AD 350–395) leading to the almost complete abandonment of the city soon after the second event [24, 25]. This abandonment and time covering favored the exceptional preservation of multiple damages affecting the main religious and city buildings: the Apollo Temple, forum, Decumanus, theater, hot-springs, etc. Also, the work of RodríguezPascua [14] defined the classification of the different archaeological earthquake damages (EAEs) by the exceptional record in Baelo Claudia, whereas in this work we introduce the building oriented damage (BOD) based on the same data.

60.3

Building Oriented Damage (BOD): Archaeological Seismic Structural Analysis

BOD is based on the geological structural analysis of brittle strain methodology of spatial data, by mapping the single data obtained from the classification of EAE affecting one city or archaeological site potentially affected by an earthquake. The method allows establishing the orientation of seismic shock direction (SSD) specifically calculated for each one of the analyzed EAEs. We have applied the BOD in the instrumental earthquake of Lorca, Spain (2011), to check the quality of the results obtained from this analysis in the historical earthquake of Baelo Claudia. This testing has been performed with the field trip carried out shortly after the Lorca earthquake (Spain) on May 11, 2011 (Mw 5.2) [27], an earthquake that killed nine people of the town, dozens were injured, almost 80% of the buildings were affected, and one modern edifice collapsed. This earthquake generated more than 144 recognized damages EAE-type in the city and affecting historic and modern buildings [15]. Measured EAEs were oriented drops of decorative elements on façades, tilted walls and collapsed structures, fallen keystones in arches, masonry block displacements, and breaking in corners of blocks (dipping broken corners). The BOD workflow combines EEEs and EAEs generated during the ground shaking, consists of two main stages: (a) damage identification and inventory of primary effects (geological and effects on buildings) and (b) strain damage quantification by measuring of oriented single data of EAEs, i.e., fallen columns, and EEEs, i.e., ground cracks and folding (Fig. 60.1).

1768

J. L. Giner-Robles et al.

60.3.1 Damage Identification and Inventory Damage identification of EAEs, inventory, and orientation is the first step. However, the main difficulty is the assignation of these effects to a single earthquake, for past events. Hence, it is mandatory for a preliminary step of searching historical documents describing the effects of damage in the archaeological site, building, or man-made structure. Overlapping damages of past earthquakes can be discriminated by checking ancient documents regarding the restorations. In the case of ancient archaeological sites with no written documents of past earthquakes, the official excavation reports can shed light to detect possible archaeoseismological effects and clues for relative dating of seismic damaged structures. Finally, any information regarding relevant archaeological restoration and/or consolidation works of ruins and buildings have to be consulted to discard the restored areas from the quantification analysis. The classification of earthquake archaeological effects (EAEs) proposed by Rodríguez-Pascua et al. [14] separates earthquake geological effects and earthquake effects on building fabric, defining several damage features which can be oriented and mapped. The archaeological and historical information must be considered to (a) identify and interpret correctly the existing EAEs with a single event and/or patterns of strain in agreement with multiple events and (b) avoid other causes which can be responsible for the damage. However, some of the inventoried effects could have not an unequivocally seismic origin, and data should not be included in the strain quantification analysis to minimize uncertainties.

60.3.2 Strain Seismic Damage Quantification The strain seismic damage quantification is based on the mapping of oriented EAEs assuming the point data as either a line (orientation) or as a plane (tilted wall). As an advantage, the obtained results can be compared with results from the analysis of earthquake environmental and geological effects (EEEs) given the similarity between both methodologies.

Data Characterization The analysis includes three main parameters: (1) data orientation, (2) type of data, and (3) data complementarity. Data orientation. The “data orientation” depends on the nature of the EAE [14] and can be grouped in two main categories: (a) data considered as a “plane” (P) and defined by measuring the plane direction and dip, for instance, folded surfaces, flagstone orientation on pavements, tilted walls, etc., and (b) data defined by the “orientation of a line” (Fig. 60.2), which can be subclassified in three line types: (L1) single line direction (e.g., axis of folds); (L2) azimuth (e.g., column fall azimuth), and (L3) a vector defined by the azimuth and magnitude (e.g. displacement vectors of block masonry defined by the slip direction and distance of displacement).

60

Geological Structural Analysis Applied to Archaeoseismology

1769

Fig. 60.2 Characterization and type of linear data: (a) L1, the direction of SSD can be defined (e.g., folds in pavements); (b) L2, the effects have clear directionality apart from establishing direction, and the azimuth of the damage can be recognized (e.g., fallen columns); (c) L3, azimuth can be obtained as in L2, but the magnitude of the deformation can also be quantified (e.g., block displacement vectors: azimuth and magnitude of that displacement). See Fig. 60.6 for color and symbol legend

Type of data. The aforementioned data orientation (either a plane P or a line L) can be defined by a unique vector orientation (O) or by an interval of orientations (R) between two directions. For example, shock breakouts in flagstones on pavements can be defined by a single vector (Fig. 60.2a). However, a range of directions fit as the potential orientation for the seismic shock direction (SSD) regarding the seismic effect of tilted walls (see Fig. 60.3a). Data complementarity. Complementarity establishes whether the analysis of different EAEs of similar types can reduce the uncertainty for determining the SSD

1770

J. L. Giner-Robles et al.

Fig. 60.3 Example of (a) range of orientations and (b) complementary data analysis. In the case of tilted or collapsed walls, a range of SSD can be identified, and measuring two walls of the same house, the SSD interval (and the uncertainty) is reduced. See Fig. 60.6 for color and symbol legend

orientation. An example is the analysis of tilted walls during the earthquake shaking. In the case of the historical earthquake of Baelo Claudia (Cadiz, Spain) with an orthogonal urban spatial arrangement of streets and buildings, we can use the walls of the same building for defining the interval of SSD intersection bearing in mind that each wall defines an azimuth interval for SSD of around 180 (see Fig. 60.3b for graphical explanation). As aforementioned, there are EAEs that provide either a single orientation (single data, Sd) or a range of orientations (complementary data, Cd). Those complementary data (walls, dipped keystones on arches, among others), must be compared with other data obtained from the analysis of the same EAE in other areas on the archaeological site, for making robust the assignment of a single paleoevent. The assignment of the three different parameters for data characterization, (1) data orientation, (2) type of data, and (3) data complementarity, leads to the second step for the quantification of the strain seismic damage analysis.

Strain Seismic Damage Analysis The analysis of the deformation figures out the orientation of the SSD related to each EAE. The nature of the ground movement during the earthquake in each EAEs determines the type of data interpreted and mapped. For example, in the case of dropped keystones in arches (Fig. 60.4), the deformation can be interpreted as the result of a repetitive horizontal movement during the earthquake, which is assumed subparallel to the wall containing the arch structure. This horizontal movement generates the separation of the contiguous keystones, eliminating their support and finally causing their fall (commonly the central keystone and annexed pieces, Fig. 60.4a). In this example, SSD could be oriented in a range of orientations (L1, R), defined by the direction of the plane which comprises the arch 45 (Fig. 60.4b). Nevertheless, we can reduce the range of orientations for the SSD orientation (and

60

Geological Structural Analysis Applied to Archaeoseismology

1771

Fig. 60.4 Simplified diagrams of strain analysis in arches and lintels. Deformation caused by seismic events induces the collapse of arch segments (a). In this case, the range of azimuths for the SSD is 45 from the plane of the arch (b). Orientations close to the perpendicular direction of the plane comprising the arch do not provoke the displacement of the central keystone. See Fig. 60.6 for color and symbol legend

the uncertainty as well), by using other EAE of similar typology (complementary data). An example of complementary data is shown in Fig. 60.3b. Three examples of different EAES from the city of Baelo Claudia and originally defined by Giner et al. [19] and Rodriguez-Pascua et al. [14] are shown in Fig. 60.5 with the type of data and their orientation. Tilted and folded walls (Fig. 60.5a, b), folds, and pop-ups on regular pavements (Fig. 60.5c) and shock breakouts (Fig. 60.5d) appear across the city, affecting coeval buildings of the same cultural period. We introduce a group of cartographic symbols for mapping in archaeological sites and modern cities affected by earthquakes (Fig. 60.6). The map of SSD allows an ata-glance interpretation of the damage directionality and pattern of seismic damage. It is recommended the use of this map in an archaeological site to show the presence of local induced geological phenomena, which can produce oriented local deformations (e.g., landslides) non-earthquake related. Besides, different external aspects can influence the damage orientations, for example, the topography where the city is placed. In this case, it is necessary to take into account the orientation of the land slopes when interpreting the results, to avoid local effects such as non-earthquakeinduced landslides. For example, it is possible to observe how the topographic slope can induce oriented deformations that can be interpreted as damage caused by a seismic event at the Roman town of Mulva-Munigia (Seville, Spain) [28]. In this case, it is necessary to complement the BOD with the archaeological and geological dating of the deformations and geomorphology study to distinguish the origin of damage (see Fig. 60.1). In some cases, the structure of the affected architectural element can determine the orientation of the damage as well. Giner-Robles et al. [29] establish an analysis of different architectural structures affected by instrumental earthquakes: Christchurch seismic sequences (New Zealand, September 2010 and February 2011); Lorca earthquake (Spain, May 2011, Fig. 60.8); and Emilia Romagna earthquake (Italy, May 2012). These authors explained the intrinsic constraints of the orientation of the seismic damage according to the nature of the EAE.

1772

J. L. Giner-Robles et al.

Fig. 60.5 Examples of strain characterization in structures of the Roman archaeological site of Baelo Claudia (Cadiz, Spain). (a, b) Tilted and folded walls (city wall), (c) folds and pop-ups on regular pavements (forum of the city), blue petals of the rose diagrams are data, and orange is the range of directions; (d) shock breakouts in flagstones (Decumanus Maximus). See Fig. 60.6 for color and symbol legend

Archaeological Site Analysis The origin of ancient deformations or levels of destruction in archaeological sites is always controversial due to a large amount of potential uncertainties. This is the reason why we suggest carrying out a combined analysis of all the data to establish the “potential seismic origin” of the deformations. This analysis combines the EAE

60

Geological Structural Analysis Applied to Archaeoseismology

1773

Fig. 60.6 Cartographic symbols and data characterization of the parameters considered in the strain seismic damage analysis of EAE. Data orientation: planes (P) or lines (L1, L2, and L3); and data type, single orientation (O) or range of orientations (R). Single data (Sd) or complementary data (Cd)

study with the classic archaeological study of urban planning, the building fabrics, and the orientation of streets or wall structures. Furthermore, the presence of other subterranean structures that can modify SSD orientations, such as underground pipes, sewage systems, and old foundations has to be considered (e.g., Giner-Robles

1774

J. L. Giner-Robles et al.

et al. [19] about the sewage under the Baelo Claudia pavement). Despite the quantitative nature of BOD and the absence of a likelihood value of uncertainty, the power of this technique is the mapping of all of EAEs for searching homogeneous and clear seismic and strain patterns. Therefore, if the results of a set of EAE are homogeneous throughout the whole archaeological site, the deformation origin could be assumed as a paleoevent or a historical earthquake (Fig. 60.7) and dated according to the age of the cultural period affected. All structures and EAEs must be properly dated, and the ages have to be consistent with the event to assign the whole damage to a single paleoevent. The map of SSD for the Lorca earthquake (Fig. 60.8) shows an example of how a single event produces a set of different EAEs compatible with a coeval shaking of different buildings and man-made urban elements. The SSD shows a NW-SE trending compatible with the value of peak ground acceleration recorded during the earthquake and the seismic surface particle motion [30].

Fig. 60.7 Results of the analysis of the archaeological site of Baelo Claudia (Cadiz, Spain). Cartographic representation of the different EAEs inventoried in the downtown area of this Roman city. Rose diagram of the SSD obtained from the analysis of the different EAEs. The range of orientations is common for the whole archaeological site, according to SSW trending and defining a single seismic event with a seismic shock direction of NE-SW trending

60

Geological Structural Analysis Applied to Archaeoseismology

1775

Fig. 60.8 (a) Map of the SSD direction obtained by the analysis of EAEs measured during the field trip after the Lorca earthquake (Mw 5.2 May 2011). (b) Accelerometer location (red square). (c) Rose diagram of SSD orientations (41 data) for the enlarged area (b). (d) Motion vectors obtained from the analysis of the movement of particle 2D registered on the accelerometer. (Modified from Giner et al. [15])

1776

60.4

J. L. Giner-Robles et al.

Conclusions

The BOD analyses may identify the potential existence of the seismic deformation affecting archaeological sites and the fact that a destructive earthquake could be the origin of the strain patterns in buildings. This analysis assume that site damage caused by an earthquake produces a specific spatial pattern of earthquake archaeological effects (EAEs) related to the seismic shock direction (SSD). Moreover, the SSD is related to the seismic source location, the epicentral depth, the landscape, and the local geology. The proposed methodology combines the inventory of EAEs with the seismic damage analysis of buildings and man-made structures and archaeological dating, by using techniques similar to those used for mapping tectonic strain typical for structural geology [19]. Therefore, the proposed approach deciphers the potential seismic origin of building damage in archaeological sites where sudden abandonment of horizons is present. The use of historical documents, official archaeological reports, and dating geological and archaeological techniques is mandatory to assign a single event in space and time. Furthermore, the combination with the analysis of environmental earthquake effects gives the total area affected by the earthquake, being possible to determine the macroseismic ESI-07 intensity value. Consequently, these archaeoseimological studies can help to incorporate unknown ancient earthquakes into the seismic catalogs and to identify probable seismic sources to improve future seismic hazard analyses. Acknowledgments This work is supported by the Spanish research project MINECO-FEDER CGL2015-67169-P (QTECSPAIN-USAL) and 3GEO, CGL2017-83931-C3-2-P. This is a contribution of the QTECT-AEQUA working group.

References 1. Karcz I, Kafri U (1978) Evaluation of supposed archaeoseismic damage in Israel. J Archaeol Sci 5:237–253. https://doi.org/10.1016/0305-4403(78)90042-0 2. Rapp G (1982) Earthquakes in the troad. In: Rapp G, Gifford JA (eds) Troy: the archaeological geology. Princeton University Press, Princeton, pp 43–58 3. Nur A, Ron H (1996) And the walls came tumbling down: earthquake history in the Holyland. In: Stiros S, Jones RE (eds) Archaeoseismology. Fitch laboratory occasional paper, vol 7. British School at Athens, Athens, pp 75–86 4. Nikonov A (1988) On the metholology of archaeoseismic research into historical monuments. In: Marinos G, Koukis G (eds) Engineering geology of ancient works, monuments and historical sites. Balkema, Roterdam, pp 1325–1320 5. Stiros S (1996) Identification of earthquakes from archaeological data: methodology, criteria and limitations. In: Stiros S, Jones R (eds) Archaeoseismology. Fitch laboratory occasional paper, vol 7. British School at Athens, Athens, pp 129–152 6. Marco S (2008) Recognition of earthquake-related damage in archaeological sites: examples from the Dead Sea fault zone. Tectonophysics 453:148–156 7. Ambraseys E (2006) Earthquakes and archaeology. J Archaeol Sci 33:1008–1016 8. Stiros S, Jones R (eds) (1996) Archaeoseismology, Fitch laboratory occasional paper, vol 7. British School at Athens, Athens. 256 pp

60

Geological Structural Analysis Applied to Archaeoseismology

1777

9. Silva PG, Reicherter K, Ch. Grützner, Bardají T, Lario J, Goy JL, Zazo C, Becker-Heidmann P (2009) Surface and subsurface palaeoseismic records at the ancient Roman city of Baelo Claudia and the Bolonia Bay area, Cádiz (south Spain). Geological Society of London, Special Publication, 316, London. https://doi.org/10.1144/SP316.6 10. Sintubin M, Stewart IS, Niemi TM, Altunel E (eds) (2010) Ancient earthquakes. Special papers, 471. Geological Society of America. https://doi.org/10.1130/SPE471 11. Ghayamghamian MR, Tobita T, Iai S, Kang G (2007) Reconnaissance Report of July 16, 2007 Niigata-Ken Chuetsu-Oki, Japan, Earthquake. Technical Note (Archive of SID), 9 (1, 2), 73 pp 12. Konagai K, Akbarl A, Oguni K, Kodama H, Ikeda T (2005) Provisional report of the damage caused by Muzaffarabad earthquake of october 8, Pakistan, 17 pp 13. Motokí K, Seo K (2000) Strong motion characteristics near the source region of ThehyogokenNanbu Earthquake from analyses of the directions of structural failures, 12WCEE, 959–965 14. Rodríguez-Pascua MA, Pérez-López R, Silva PG, Giner-Robles JL, Garduño-Monroy VH, Reicherter K (2011) A comprehensive classification of earthquake archaeological effects (EAE) for archaeoseismology. Quat Int 242:20–30. https://doi.org/10.1016/j.quaint.2011.04.044 15. Giner-Robles JL, Pérez-López R, Silva PG, Rodríguez-Pascua MA, Martín-González F, Cabañas L (2012) Análisis estructural de danos orientados en el terremoto de Lorca del 11 de mayo de 2011. Aplicaciones en Arqueosismología. Bol Geol Min 123(4):503–513 16. Reches Z (1983) Faulting of rocks in three-dimesnional strain fieldsII. Theoretical analysis. Tectonophysics 95:133–156. https://doi.org/10.1016/0040-1951(83)90264-0 17. De Vicente G (1987) The e/k’ diagram. An application of the slip model to the population fault analysis. Rev Soc Geol Esp 1:97–112 18. Michetti AM, Esposito E, Guerrieri L, Porfido S, Serva L, Tatevossian R, Vittori E, Audemard F, Azuma T, Clague J, Comerci V, Gurpinar A, Calpin JM, Mohammadioun B, Morner NA, Ota Y, Roghozin E (2007) In: Guerrieri L, Vittori E (eds) Intensity scale ESI 2007. Memorie Descrittive Carta Geologica d’Italia, vol 74. Servizio Geologico d’Italia – Dipartimento Difesa del Suolo, APAT, Roma 19. Giner-Robles JL, Rodríguez-Pascua MA, Pérez-López R, Silva PG, Bardají T, Grützner C, Reicherter K (2009) Structural analysis of earthquake archaeological effects (EAE): Baelo Claudia Examples (Cádiz, South Spain). Instituto Geológico y Minero de España, Madrid 20. Rodríguez-Pascua MA, Silva PG, Pérez-López R, Giner-Robles JL, Martín-González F, Perucha MA (2013) Preliminary intensity correlation between macroseismic scales (ESI07 and EMS98) and Earthquake archaeological effects (EAEs). In: Seismic hazard, critical facilities and slow active faults. RWTH Aachen University, Aachen, pp 221–224 21. Silva PG, Elez J, Giner-Robles JL, Rodríguez-Pascua MA, Pérez-Lóez R, Roquero E, Bardají T, Martínez-Graña A (2016) ESI-07 ShakeMaps for instrumental and historical events in the Betic Cordillera (SE Spain): an approach based on geological data and applied to seismic hazard. Quat Int 451:185–208. https://doi.org/10.1016/j.quaint.2016.10.020 22. Worden CB, Wald DJ (2016) ShakeMap manual online: technical manual, user’s guide, and software guide. U. S. Geological Survey. usgs.github.io/shakemap. https://doi.org/10.5066/ F7D21VPQ 23. Silva PG, Rodríguez-Pascua MA, Giner-Robles JL, Élez J, Huerta Hurtado P, García Tortosa F, Bardají Azcárate T, Perucha Atienza MA, Vicente Gómez P, Pérez-López R, Lario Gómez J, Roquero García-Casal E, Bautista Davila MB (2019) In: Silva PG, Rodríguez-Pascua MA (eds) Catálogo de los efectos geológicos de los terremotos en España, 2nd edn. Instituto Geológico y Minero de España, IGME, Madrid 24. Menanteau L, Vanney JR, Zazo C (1983) Belo II: Belo et son environment (Detroit de Gibraltar). Etude physique d’un site antique. Casa de Velázquez. Serie Arqcheologie, 4, Broccard, París 25. Sillières P (1997) Baelo Claudia: Una ciudad Romana de la Bética. Junta de Andalucía- Casa de Velázquez, Madrid 26. Silva PG, Borja F, Zazo C, Goy JL, Bardají T, De Luque L, Lario J, Dabrio CJ (2005) Archaeoseismic record at the ancient Roman City of Baelo Claudia (Cádiz, south Spain). Tectonophysics 408(1–4):129–146. https://doi.org/10.1016/j.tecto.2005.05.031

1778

J. L. Giner-Robles et al.

27. Martínez-Díaz JJ, Béjar-Pizarro M, Álvarez-Gómez JA, Man-cilla FL, Stich D, Herrera G, Morales J (2012) Tectonic and seismic implications of an intersegment rupture. The damaging May 11 th 2011 Mw 5.2 Lorca, Spain, earthquake. Tectonophysics 546–547. https://doi.org/10. 1016/j.tecto.2012.04.010 28. Giner-Robles JL, Bardají T, Rodríguez-Pascua MA, Silva PG, Roquero E, Elez J, Perucha MA, Baena R, Guerrero I, Fernández-Caro JJ, Pérez-López R, Rodríguez-Escudero E (2016) Archaeoseismological analysis of Mulva-Munigua roman archaeological site (Sevilla, Spain). Preliminary results. Geo-Temas 16(1):605–608 29. Giner-Robles JL, Pérez-López R, Rodríguez-Pascua MA, Silva PG, Martín-González F, Rodríguez-Escudero E (2013) A review on oriented fall structures (Earthquake Archaeological Effect, EAE) induced by instrumental earthquakes. Cuaternario y Geomorfología 27(3–4):5–32 30. Cabañas L, Alcalde JM, Carreño E, Bravo JB (2013) Characteristics of observed strong ground motion accelerograms from the 2011 Lorca (Spain) earthquake. Bull Earthq Eng 12:1909–1932. https://doi.org/10.1007/s10518-013-9501-0 31. Giner-Robles JL, Rodríguez-Pascua MA, Silva PG, Pérez-López R (2018) Efectos sísmicos en yacimientos arqueológicos: catalogación y cuantificación arqueosismológica. Bol Geol Min 129(1–2):451–467. https://doi.org/10.21701/bolgeomin.129.1.018

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic Damage and Intensity Assessments in the Cultural Heritage Miguel A´ngel Rodríguez-Pascua, Pablo Gabriel Silva, Jorge Luis Giner-Robles, María A´ngeles Perucha, Elvira Roquero, Teresa Bardají, Javier Elez, and Rau´l Pe´rez-Lo´pez Contents 61.1 61.2 61.3 61.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Environmental Seismic Intensity Scale (ESI-07) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . European Macroseismic Scale EMS-98 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Example Case of the Ancient Roman City of Baelo Claudia (Cádiz, Spain) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1780 1782 1782 1784 1786 1788

Abstract

The earthquake archaeological effects (EAEs) and their geological structural analysis can be used to determine or reject the seismic origin of deformations in historical buildings or ancient structures. But, nowadays it is not possible to

M. Á. Rodríguez-Pascua (*) · M. Á. Perucha · R. Pérez-López Instituto Geológico y Minero de España (IGME), Madrid, Spain e-mail: [email protected]; [email protected]; [email protected] P. G. Silva · J. Elez Dpto. Geología, Universidad de Salamanca, Avila, Spain e-mail: [email protected]; [email protected] J. L. Giner-Robles Dpto. Geología y Geoquímica. Fac. Ciencias, Universidad Autónoma de Madrid (UAM), Madrid, Spain e-mail: [email protected] E. Roquero Dpto. de Edafología. E.T.S.I. Agrónomos, Universidad Politécnica (UPM), Madrid, Spain e-mail: [email protected] T. Bardají U.D. Geología, Universidad de Alcalá (UAH), Madrid, Spain e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_61

1779

M. A´. Rodríguez-Pascua et al.

1780

define the seismic intensity for particular EAEs. Using the main constructive styles previous to the twentieth century, especially those related to the historical heritage, this paper proposes a preliminary correlation between the different EAEs with the macroseismic scales EMS-98 and ESI-07. We propose minimum and maximum intensity levels for different building types, such as those related to the use of adobe, stone, brick, and masonry. This work aims to establish preliminary seismic intensity intervals from the nature and styles of seismic deformations in historical buildings that can be applied to the archaeoseismological analysis of heritage buildings and archaeological sites.

61.1

Introduction

Earthquakes that took place during historical times can produce several types of deformational structures that can be preserved in archaeological sites and heritage buildings. These, called earthquake archaeological effects (EAEs) [1], contribute to refine the seismic history of a particular region and constitute an interesting seismic record preserved in the cultural heritage that can be used in seismic prevention. However, at present still does not exist any specific macroseismic scale for intensity assessments from the different EAEs occurring in our historical heritage. The EMS-98 [2] is the official harmonized macroseismic scale in the European Union, and this intensity scale does not consider seismic damage of historical buildings for the evaluation of the intensity in a locality. The EMS-98 excludes also the earthquake environmental effects (EEEs) that in many cases largely contribute to producing important damages in the cultural heritage during earthquakes. The macroseismic scale ESI-07 [3] was created to complement and mitigate the absence of intensity assessments from EEEs in the EMS-98 scale [4]. In addition, the ESI-07 scale can be applied for different historical periods and in depopulated areas, covering the entire territory not only urbanized zones in the cities [5]. For this reason, using the classification of EAEs proposed by [1] this study proposes a correlation between the EAEs and the EMS-98 and ESI-07 macroseismic scales. The proposed correlation is based on the occurrence of different types of EAEs in the building fabric at different EMS-98 intensity levels and the related earthquake environmental effects (EEEs) cataloged in the ESI-07 scale [6]. The idea is to provide a methodology leading to estimate the seismic intensity from archaeoseismological analyses in the historical heritage: an archaeological macroseismic scale based on the analysis of EAEs. The base of this new macroseismic scale will be the classification chart of EAEs published by [1] (Fig. 61.1). This chart differentiates primary from secondary earthquake effects on buildings. The secondary effects occur after the earthquake (post-seismic effects) and are excluded from the proposed macroseismic scale. On the contrary, the primary effects are related to the coseismic disturbances and are subdivided into two categories: (a) geological effects on the ground and (b) effects in the building fabric (deformations).

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic. . .

1781

Fig. 61.1 Earthquake archaeological effects (EAE) classification relative to intensity intervals in which these structures can be generated taking into account the macroseismic scales ESI07 and EMS98. The black bars indicate the minimum intensity limit from which is generated a certain EAE. (Modified from Rodríguez-Pascua et al. Ref. [1])

The geological effects of the chart are included in the macroseismic scale ESI-07 and correspond to the different types of EEEs considered in this scale. For this reason, we choose the intensity thresholds of the ESI-07 for their application to the EAEs when they appear related to ground geological effects (EEEs). On the other hand, the EMS-98 scale considers soil liquefaction, landslides, rockfalls, and compacted anthropic substratum as probable ground failures of geological origin that could affect building damage. In this way, we can compare particular intensity thresholds of the ESI-07 and EMS-98 scales for the occurrence of individual EAEs in different constructive styles (Fig. 61.1). Effects in the building fabric are fully included by the EMS-98 scale and can be used to set the limit of the minimum intensity that can give rise a certain EAE. The total destruction of the building implies that the intensity scale saturates above this limit and represents the maximum intensity for all of the EAEs.

M. A´. Rodríguez-Pascua et al.

1782

61.2

Environmental Seismic Intensity Scale (ESI-07)

The geological effects were seriously dismissed in the structure of the EMS-98 intensity scale, but largely differentiated and parameterized by the ESI-07 one [3]. This last is used here to assign the intensity intervals to the EAEs related to geological effects, as can be liquefactions, slope movements, etc. [6]. One of the most important features of the ESI-07 scale is that it allows to define seismic intensity for the entire affected area independently of the existence of cities, towns, or villages and, if the case, only in the county side. This is obviously quite important for intensity assessments in archaeological sites. The ESI-07 scale integrates the traditional intensity scales that consider EEEs and can be used combined with different macroseismic scales, including the EMS-98, to generate hybrid seismic scenarios [3], which is useful to define archaeoseismic intensities proposed in this work. All the geological effects at their most energetic level can produce the total destruction of an archaeological site or a heritage building. Consequently, the maximum intensity level considered in this work will always be XII for all the EAEs. The minimum limit is assigned using the lowest intensity at which a geological effect can be preserved in the geological or geo-archaeological record. In the case of landslides, rockfalls, anthropic compacted ground, and liquefactions are possible to establish these minimum bounds by means of the combination of both macroseismic scales (ESI07 and EMS98) as noticed by [6]. Some intensity discrepancies in the lower bounds of these scales only occur for anthropic compacted ground and rockfalls. For the rockfalls, the minimum EMS-98 intensity is VI, but IV for the ESI-07 intensity (Fig. 61.1), although for those lower intensities, rockfalls and small slope movements commonly occur in artificial or modified talus.

61.3

European Macroseismic Scale EMS-98

In the case of the establishment of the minimum values of intensities for the EAEs on the fabric of buildings, we only used the EMS-98 scale. To establish the minimum values, all the construction styles listed in the EMS-98 scale were considered, which in turn are normally used in historical heritage buildings and archaeological sites: (1) rubble stone; (2) adobe (earth brick); (3) simple stone; (4) brick; and (5) masonry. Comparing these five constructive types with the vulnerability intervals (five values in alphabetical code: A–F) and the building damage classes (numerical code: 1–5) considered in the EMS98 scale, it is possible to establish the minimum intensity values for which damages (EAEs) appear (Fig. 61.2). The most vulnerable constructive types have intensity V (rubble stone and adobe), but the most seismic-resistant constructive types (masonry) reach this minimum value at intensity VII (Fig. 61.3). The total destruction of these constructive types is considered to occur from intensity VIII to IX (rubble stone and adobe) in the EMS-98 but reaches intensity X for the more resistant masonry buildings (Fig. 61.3). Above intensity X destruction effects saturate for the considered constructive types, and

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic. . .

1783

Fig. 61.2 Seismic intensity calculation process using the EMS98 scale for the development of conjugated fractures in walls made of bricks. (Modified from Ref. [2])

Fig. 61.3 EMS98 seismic intensity intervals from the damage onset and destruction in historical buildings and archaeological sites. Also the vulnerability values and damage classes are considered for fabric buildings proposed by this macroseismic scale

only building damage in metallic-structure and seism-resistant modern edifices (no existing in historical times) can be used for intensity assessments. Between these minimum and maximum values, there is a zone in which all EAEs occur, being possible to establish indicative seismic intensity intervals for particular EAEs before the total destruction of a building [6]. This is the way to establish

1784

M. A´. Rodríguez-Pascua et al.

the intensity values associated with each EAE. The first step is to select the constructive type and the type of associated structure (e.g., walls made of bricks), from which we can set the assigned vulnerability class in the EMS-98 scale (e.g., A–C). In a second step, the lower and higher limit of intensity for the analyzed EAE is obtained from the building damage classes (numerical code) associated to this effect (Fig. 61.2). Figure 61.1 displays how the vulnerability and the minimum damage classes are calculated for every EAE. The obtained values are the key code to obtain the minimum and maximum intensity that can give rise to a particular EAE. Combining the intensity intervals extracted from the ESI-07 scale (Fig. 61.1) and the EMS-98 scale (Fig. 61.2), the EAE lower minimum value is intensity IV (rockfalls) and the EAE higher minimum value intensity X (tsunamis). Above these values the scale will be saturated (I ¼ VIII–X from rubble stone to masonry; Fig. 61.3). If a particular archaeological site/building records different EAEs, these deformation structures may indicate the minimum and maximum values of the “bracketed seismic intensity” affecting the site. These bracketed intensity values (minimum and maximum) will indicate the seismic intensity interval in which damage was produced (e.g., VII–X). Depending on the frequency and size of the maximum intensity EAE indicators finer assessments are possible, and double (e.g., VIII–IX) or unique (e.g., X) intensity values can be proposed. In any case, the use of the proposed methodology is particularly interesting and recommended “when and where” EAE indicators will be closely linked to earthquake environmental effects (EEEs). On the contrary, their application is inadvisable for those buildings or archaeological sites with few EAEs.

61.4

The Example Case of the Ancient Roman City of Baelo Claudia (Ca´diz, Spain)

The ancient Roman city of Baelo Claudia is a well-known archaeoseismological site located in Bolonia Bay (Cádiz, SW of Spain), and it is credited to be the subject of the first archaeoseismological studies in the Western Mediterranean [7]. Presently, the Roman remains are protected and since the 1980s decade is within the network of archaeological parks of the Andalusia Regional Government. Baelo Claudia was founded in the second century BC and was destroyed two times by earthquakes (AD 40–60 and AD 350–395). After the second earthquake, the city was abandoned [8, 9]. The destruction layers in the archaeological site were interpreted like earthquakes by these authors. The extension of the excavated zone offers hundreds of archaeoseismic data (EAEs) allowing the analysis and mapping of the distribution of EAEs along the entire archaeological site for the second earthquake in AD 350–395 [10]. The amount of analyzed EAEs made possible the estimation of the seismic intensity that destroyed the Roman city for the second time giving place to their eventual abandonment. After the abandonment of the city, the place was barely occupied by small Visigothic or Islamic settlements, and the Roman remains were preserved in excellent conditions, including the EAEs, until the

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic. . .

1785

first excavation works in the early twentieth century and the present [7]. The two ancient events recorded in Baelo Claudia are included in the new official catalog of the earthquake geological effects in Spain [11]. [7, 10], and [11] studied the deformation structures located within this archaeological site as well as their relationships with the tectonic structures of the area and their impact on the landscape. These works support the seismic hypothesis for the destruction of the city by means of quantitative approaches validating the first qualitative observations made by [8]. These modern studies provided the first catalog of preserved earthquake effects in the archaeological site, performed a first structural analysis on oriented damage, identified potential seismic sources in the near-field, and proposed preliminary intensity levels for the ancient earthquakes. Further geophysical research and fault trenching in the zone [12] helped to refine the seismic history associated to the Roman city. Using the earthquake effects preserved in the Roman remains of Baelo Claudia, [13–16] launched an innovative approach by means of the systematic structural geology analysis of the EAEs to identify oriented damage of seismic origin. These authors cataloged more than 100 EAEs to eventually produce a detailed map of the different damaged sectors of the Roman city during the last earthquake in 350–395 AD [13–15]. The methodological approach proposed by these authors was eventually focused on the estimation of the mean direction of ground motion during the earthquake. They obtained consistent NE-SW mean orientation of ground motion using more than 100 different EAEs, congruent with a near-field offshore seismic sources SSW of the Bolonia Bay. Further seismic profiling in the offshore zone identified at least a couple of NNW-SSE active faults affecting the Plio-Quaternary filling of the bay and the submarine topography [12]. Considering the complete sets of EAEs cataloged by the aforementioned authors, the most important EAEs listed by [1] are present in the ancient city of Baelo Claudia. This fact makes this archaeological site an important locality (maybe unique) for the evaluation of bracketed intensity levels by means of the hybrid ESI-EMS methodology proposed in this work. The most common EAEs recorded in this site are displaced masonry blocks (Fig. 61.4) and dipping broken corners (Fig. 61.5), and the estimated seismic hybrid intensity intervals obtained by the combination of the EMS-98 and ESI-07 scales are the following ones for the whole of the EAEs (Fig. 61.6): Analyzing the EAEs recorded in this archaeological site (Table 61.1) is clear that ground shaking exceeded intensity VIII during the AD 350–395 earthquake. Taken into account the EAE markers with higher minimum intensity values (i.e., DMB; Fig. 61.4; Table 61.1), a minimum intensity of VIII can be firmly proposed for this event, but in this range, the earthquake could reach intensity X or higher (difficult to assess). Authors like [8] or [9] suggested an X MSK intensity is considered as a maximum value for [10, 11], which considers a range of VIII–X intensity. Nevertheless, considering the frequency and size (length or amplitude) of all the EAEs markers of intensity VII and VIII, as well as the high number of centimeter-scale DMB recorded in the site, Intensity VIII–IX is the minimum bracketed value to be

M. A´. Rodríguez-Pascua et al.

1786 Fig. 61.4 Displaced masonry blocks in the market of the ancient Roman city of Baelo Claudia. Minimum macroseismic EAE intensity: IX

considered for this archaeological site. On the other hand, the FWL has a minimum intensity of VIII (Table 61.1) and will be the higher minimum value for EAEs. Consequently, the interval of intensity for this earthquake will be VIII–X (Fig. 61.6).

61.5

Conclusions

The proposed methodology for the estimation of macroseismic intensities in archaeological and/or historic buildings is based on the combination of the macroseismic scales ESI-07 and EMS-98, as well as in the classification of EAEs of Rodriguez-Pascua et al. (2011). From the intensity ranges proposed by both scales, it is possible to set the minimum intensities values that can generate several EAEs. This makes it possible to estimate the intensity intervals triggering the different EAEs. In the case of Baelo Claudia, the number of cataloged EAEs, their overall measured orientation, their frequency, and their size allow the application of the proposed methodology. Its application is possible only with a numerous set of

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic. . .

1787

Fig. 61.5 Dipping broken corners in the walls of the market of the ancient Roman city of Baelo Claudia made of masonry blocks. Minimum macroseismic EAE intensity: VII

particular EAEs or with few sets of different EAEs. Their application to isolated cases is not recommended. From this analysis, it is clear that intensity VII is the minimum value from which different sets of EAEs become frequent but widespread and diagnostic from intensities  VIII. Taking into account that several of the buildings of Baelo Claudia collapsed during the event (i.e., basilica, market, temples) and the severe damage caused in the pavement of the Decumanus maximums, forum, and theater [9] and related ground effects [11], a minimum “archaeoseismic intensity” in the range of IX–X can be firmly proposed for the AD 350–395 event [17]. The archaeoseismological methodology proposed in this work offers bracketed intensity values indicating the minimum intensity value recorded in the historical heritage and archaeological sites. This is a preliminary step to the systematization and further parametrization of EAEs. Acknowledgements This work is supported by the Spanish research project MINECO-FEDER CGL2015-67169-P (QTECSPAIN-USAL). This is a contribution of the QTECT-EQUA working group.

M. A´. Rodríguez-Pascua et al.

1788

Fig. 61.6 Intensity intervals for the EAEs observed in Baelo Claudia. The yellow rectangle marks the intensity interval for the calculated intensity: VIII–X Table 61.1 Set of EAEs types recorded in the Roman archaeological site of Baelo Claudia (Cádiz, South Spain) Code IBM DBC PFB FSK SBK TWA FOC FPV DWL FWL DMB

EAE Impact block marks Dipping broken corners Penetrative fractures masonry blocks Folded steps and kerbs Shocks breakouts in flagstones Tilted walls Fallen oriented columns Folded pavements Displaced walls Folded walls Displaced masonry blocks

Bracketed intensity VI–XII VII–X (saturated) VII–XII VII–XII VII–XII VII–XII (ESI; geological record) VIII–XII (ESI; geological record) VIII–XIII (ESI; geological record) VIII–XIII (ESI; geological record) VIII–XIII (ESI; geological record) IX–XII

Scale EMS-98 EMS-98 EMS-98 EMS-98 EMS-98 ESI-07 ESI-07 ESI-07 ESI-07 EMS-98 EMS-98

References 1. Rodríguez-Pascua MA, Pérez-López R, Silva PG, Giner-Robles JL, Garduño-Monroy VH, Reicherter K (2011) A comprehensive classification of earthquake archaeological effects (EAE) for archaeoseismology. Quaternary Int 242:20–30 2. Grünthal G (1998) European macroseismic scale 1998: EMS98. Musée Natioal d’Historie Narurelle, Luxembourg

61

Earthquake Archaeological Effects (EAEs) for Identification of Seismic. . .

1789

3. Michetti AM, Esposito E, Guerrieri L, Porfido S, Serva L, Tatevossian R, Vittori E, Audemard F, Azuma T, Clague J, Comerci V, Gurpinar A, Calpin JM, Mohammadioun B, Morner NA, Ota Y, Roghozin E (2007) In: Guerrieri L, Vittori E (eds) Intensity scale ESI 2007. Memorie Descrittive Carta Geologica d’Italia., 74. Servizio Geologico d’Italia – Dipartimento Difesa del Suolo, APAT, Roma 4. Serva L, Vittori E, Comerci V, Esposito E, Guerrieri L, Michetti AM, Mohammadioun B, Mohammadioun G, Porfido S, Tatevossian R (2016) Earthquake hazard and the environmental seismic intensity (ESI) scale. Pure Appl Geophys 173(5):1479–1555 5. Silva PG, Guerrieri L, Michetti AM (2015) Intensity scale ESI 2007 for assessing earthquake intensities. In: Beer M, Kougioumtzoglou IA, Patelli E, Au SK (eds) Encyclopedia of earthquake engineering. Springer, Berlin, pp 1219–1237 6. Rodríguez-Pascua MA, Silva PG, Pérez-López R, Giner-Robles JL, Martín-González F, Perucha MA (2013) Preliminary intensity correlation between macroseismic scales (ESI07 and EMS98) and Earthquake archaeological effects (EAEs). In: Grützner C, Rudersdorf A, Pérez-López R, Reicherter K (eds) Seismic hazard, critical facilities and slow active faults. RWTH Aachen University, Aachen, pp 221–224 7. Silva PG, Borja F, Zazo C, Goy JL, Bardají T, De Luque L, Lario J, Dabrio CJ (2005) Archaeoseismic record at the ancient Roman City of Baelo Claudia (Cádiz, south Spain). Tectonophysics 408(1–4):129–146 8. Menanteau L, Vanney JR, Zazo C (1983) Belo II: Belo et son environment (Detroit de Gibraltar). Etude physique d’un site antique. Casa de Velázquez. Serie Arqcheologie, 4, Broccard, París, in French 9. Sillières P (1997) Baelo Claudia: Una ciudad Romana de la Bética. Junta de Andalucía- Casa de Velázquez, Madrid, in Spanish 10. Silva PG, Giner-Robles JL, Reicherter K, Rodríguez-Pascua MA, Grützner C, Jiménez IG, García PC, Bardají T, Santos G, Roquero E, Röth J, Perucha MA, Pérez-López R, Fernández Macarro B, Martínez-Graña A, Goy JL, Zazo C (2016) Los terremotos antiguos del conjunto arqueológico romano de Baelo Claudia (Cádiz, Sur de España): Quince años de investigación arqueosismológica. Estudios Geológicos 72(1):1–24. in Spanish 11. Silva PG, Reicherter K, Ch. Grützner, Bardají T, Lario J, Goy JL, Zazo C, Becker-Heidmann P. (2009) Surface and subsurface palaeoseismic records at the ancient Roman city of Baelo Claudia and the Bolonia Bay area, Cádiz (south Spain). Geological Society of London, Special publication, 316, London 12. Grützner C, Reicherter K, Hübscher C, Silva PG (2012) Active faulting and neotectonics in the Baelo Claudia area, Campo de Gibraltar (southern Spain). Tectonophysics 554–557:127–142 13. Giner-Robles JL, Rodríguez-Pascua MA, Pérez-López R, Silva PG, Bardají T, Grützner C, Reicherter K (2009) Structural analysis of earthquake archaeological effects (EAE): Baelo Claudia Examples (Cádiz, South Spain). Instituto Geológico y Minero de España, Madrid 14. Giner-Robles JL, Pérez-López R, Silva PG, Rodríguez-Pascua MA, Martín-González F, Cabañas L (2012) Análisis estructural de danos orientados en el terremoto de Lorca del 11 de mayo de 2011. Aplicaciones en Arqueosismología. Boletín Geológico y Minero 123(4):503–513 15. Giner-Robles JL, Rodríguez-Pascua MA, Silva PG, Pérez-López R (2018) Efectos sísmicos en yacimientos arqueológicos: catalogación y cuantificación arqueosismológica. Boletín Geológico y Minero 129(1/2):451–467 16. Giner-Robles JL, Rodríguez-Pascua MA, Silva PG, Roquero E, Pérez-López R, Bardají T, Elez J, Perucha MA (2022) Methodological approach for the use of geological structural analysis applied to archaeoseismology. In: S. D’Amico and V. Venuti (eds) Cultural heritage analysis. Springer (this volume) 17. Silva PG, Rodríguez-Pascua MA, Giner-Robles JL, Élez J, Huerta Hurtado P, Tortosa FG, Bardají Azcárate T, Perucha Atienza MA, Vicente Gómez P, Pérez-López R, Lario Gómez J, Roquero García-Casal E, Bautista Davila MB (2019) In: Silva PG, Rodríguez-Pascua MA (eds) Catálogo de los efectos geológicos de los terremotos en España, 2nd edn. Instituto Geológico y Minero de España, IGME, Madrid

Cleaning of Masonry Surfaces of Cultural Interest

62

Fabio Fratini, Manuela Mattone, and Silvia Rescic

Contents 62.1 62.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Cleaning Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.1 Water Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.2 Chemical Cleaners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.3 Mechanical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.4 Laser Cleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.5 Biocleaning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.6 Environmental Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2.7 Safety Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3 Methods and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.3.1 Color Measurements: Spectrophotometric Methods . . . . . . . . . . . . . . . . . . . . 62.3.2 Superficial Cohesion Measurements: Peeling Test . . . . . . . . . . . . . . . . . . . . . . 62.3.3 Water Absorption Measurements: Contact Sponge Method . . . . . . . . . . . . 62.3.4 Aspect of the Surface: Photographic Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1792 1793 1793 1794 1795 1795 1796 1796 1796 1796 1796 1797 1798 1798 1798 1800 1803

Abstract

The contribution deals with the architectural heritage which, also not presenting monumental characteristics, bears witness of the history of each country where the building is sited. This heritage, consisting in what Roberto Pane used to define F. Fratini ISPC – CNR di Firenze, Florence, Italy e-mail: [email protected] M. Mattone Department of Architecture and Design, Polihtecnic of Turin, Turin, Italy e-mail: [email protected] S. Rescic (*) CNR-Institute of Heritage Science, Florence, Italy e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_62

1791

1792

F. Fratini et al.

as “architecture without architects,” during the last years has partially undergone ordinary maintenance works – often promoted by specific laws allowing the partial tax deduction of expenses incurred for interventions. The prevalence of economic assessments on conservative instances has determined and still determines the choice of intervention methods and techniques often not respectful enough of the buildings’ materials. This contribution intends to investigate, through portable quick-to-use instrumentation (such as camera, colorimeter, “peeling test,” water absorption by contact sponge), the outcomes of a cleaning operation, conducted on masonry walls of a civil building erected in Turin at the beginning of the twentieth century, assessing its impact in terms of material conservation.

62.1

Introduction

In the restoration of the protected monumental heritage, cleaning is generally the phase that precedes the subsequent conservative phases (protection and/or consolidation), and it is a very delicate moment because it could remove, together with the dangerous superficial deposits, even ancient treatments (e.g., oxalate patinas) and residual of chromatic decorations. Furthermore, this intervention can increase the roughness and porosity of the surface, increasing its propensity to deterioration. For this reason, in monuments, maximum attention is paid to this phase by choosing the most suitable techniques and modulating their action. This care in operations is not followed in the case of unprotected building heritage which nevertheless is the heritage that really characterizes a territory because it is wide diffusion [1–7]. This heritage, during the last years, has partially undergone ordinary maintenance works – often promoted by specific laws allowing the partial tax deduction of expenses incurred for interventions. The prevalence of economic assessments on conservative instances has determined and still determines the selection of intervention methods and techniques often not respectful enough of the building materials. As a matter of fact, these procedures are carried out without the guidance and supervision of an architectural conservator and may result in irrevocable damage to the historic resource. Just to mention, applying wrong cleaning agents to historic masonry can have disastrous results: e.g., acidic cleaners cause etching and dissolution of carbonatic stones; even on silicatic sandstones, the calcareous cement can be severely dissolved, and iron particles may oxidize and result in staining. Building elements vulnerable to deterioration may even not be visible, such as embedded ends of iron window bars, iron cramps, or ties which hold the masonry to the structural frame. The only way to prevent problems is to study the building construction in detail considering the materials of the façade (nature and condition of conservation), its structure, the kind of “dirt” (nature and composition) in order to identify the most effective and least invasive cleaning technique. On the other hand, the knowledge of the materials present in a façade is important to avoid interventions that completely distort a masonry causing the loss of the original materials desired by the designers: we want to refer to the masonries in artificial stone, a material that was in fashion at

62

Cleaning of Masonry Surfaces of Cultural Interest

1793

the beginning of the twentieth century and which is sometimes completely painted in the intervention of renovation of the façades. These are interventions that show the complete ignorance of the clients and the designers of the intervention. However, even before undertaking the study of the façade, one must ask what is the motivation for a cleaning intervention [2]: – Improve the appearance of the building by removing unattractive dirt or soiling materials or non-historic paint from the masonry. – Delay deterioration by removing soiling materials that may be damaging the masonry. – Provide a clean surface to accurately match repointing mortars or patching compounds or to conduct a condition survey of the masonry. Once clarified the aims and the result to be obtained, an exhaustive study of the façade will allow to proceed to the selection of the most appropriate technique/s also taking advantage of the execution of test areas to verify the effect obtained both as regards the aesthetic result and for possible damage caused to the surface by the technique itself. The text will first describe the different cleaning methods currently available for use on building façades, together with the criteria to be adopted for their appropriate selection. We will then discuss a cleaning operation carried out on the brick façade of a civil building erected in Turin (Italy) at the beginning of the twentieth century, investigated with portable quick-to-use instrumentations (such as camera, colorimeter, “peeling test,” water absorption by contact sponge). The cleaning of the façade with exposed bricks was carried out by hot water cleaning (pressure 120 atmospheres) without using detergents or abrasives. The idea is to assess the impact of this intervention in terms of aesthetic result and possible danger for material conservation. Nevertheless, we would like to suggest an easy procedure that, without the support of a diagnostic campaign, could help in planning a correct cleaning intervention, a procedure that can be applied by any company that deals with this type of intervention, without any additional cost for the customer.

62.2

Cleaning Methods

Masonry cleaning methods generally are divided into three major groups: water, chemical, and “mechanical” [2, 5].

62.2.1 Water Methods Water methods soften the dirt or soiling material and rinse the deposits from the masonry surface. These methods are generally the gentlest means possible, and they

1794

F. Fratini et al.

can be used safely to remove dirt from all types of historic masonry (except in the case of badly deteriorated masonry because water may exacerbate the deterioration, or in presence of soluble salts that could be mobilized, so penetrating deeply in the porosity of the masonry). Once water cleaning has been completed, it is often necessary to follow up with a water rinse to wash off the loosened soiling material from the masonry. There are essentially three kinds of water-based methods: – Soaking: prolonged spraying or misting with water. It is a very slow but also a very gentle method to be used on historic masonry. – Pressure water washing with low (100 psi or below) or medium pressure of water (generally no higher than 300–400 psi): in order to increase the capacity of this method in removing oily dirt, a nonionic detergent or surfactant can be added to the water. Adding a nonionic detergent and scrubbing with a natural bristle or synthetic bristle brush can facilitate cleaning in textured or intricately carved masonry. This should be followed by a final water rinse. – Hot-pressurized water cleaning (steam): this is a gentle and effective method for cleaning stone that can be especially useful in removing built-up soiling deposits in carved stone details. – Before applying a water cleaning method, it is important to make sure that all mortar joints are sound and that the building is watertight. Otherwise, water can seep from the surface to the interior of the masonry favoring rusting of metal anchors and staining [2, 5]. Water methods should not be used in cold weather conditions because of the possibility of freezing. Moreover, it is imperative to be aware that using water at too high a pressure (a practice common to “power washing” and “water blasting”) is very dangerous because of erosion of soft stones, as well as some types of brick and even marble, can be produced.

62.2.2 Chemical Cleaners Chemical cleaners react with dirt, soiling material, or paint giving rise to products that can be rinsed off from the masonry surface with water. Chemical cleaners used to remove dirt and soiling include acids, alkalis, and organic compounds applied with different mediums like cellulose poultice sepiolite/attapulgite poultice. Another chemical method is the application of the suitably modified gel. Acidic cleaners, of course, should not be used on masonries that are acid sensitive. Paint removers are alkaline, based on organic solvents or other chemicals. Both alkaline and acidic cleaning treatments include the use of water. Both cleaners are also likely to contain surfactants (wetting agents) that facilitate the chemical reaction that removes the dirt [8].

62

Cleaning of Masonry Surfaces of Cultural Interest

1795

More than one application of the cleaner may be necessary, and it is always a good practice to test the product manufacturer’s recommendations concerning dilution rates and dwell times. Because each cleaning situation is unique, dilution rates and dwell times can vary considerably. The masonry surface may be scrubbed lightly with natural or synthetic bristle brushes prior to rinsing. After rinsing, pH strips should be applied to the surface to ensure that the masonry has been neutralized completely. Like water methods, they should not be used in cold weather conditions because of the possibility of freezing. Both acidic and alkaline cleaners can be dangerous to cleaning operators, and clearly, there are environmental concerns associated with the use of chemical cleaners.

62.2.3 Mechanical Methods Mechanical methods include blasting with grit and the use of grinders and abrasive brushes, all of which mechanically remove the dirt, soiling material, or paint. Abrasive cleaning is also often followed by a water rinse. Generally, abrasive cleaning methods are not appropriate for use on historic masonry buildings. Since the abrasives do not differentiate between the dirt and the masonry, they also remove the outer surface of the masonry at the same time, resulting in permanently damaging the masonry. Brick, architectural terra cotta, soft stone, detailed carvings, and polished surfaces are especially susceptible to physical and aesthetical damage by mechanical methods because the abrasive powder produce a rough surface which tends to hold dirt and makes future cleaning more difficult. Abrasive cleaning processes could also subsurface micro-cracking. Abrasion of carved details causes a rounding of sharp corners and other loss of delicate features, while abrasion of polished surfaces removes the polished finish of stone. Mortar joints can also be eroded [2, 5, 9, 10]. Blasting with abrasive grit or other abrasive materials is the most frequently used cleaning method in ordinary buildings, and sand is the most common abrasive product. Finely ground silica or glass powder, glass beads, ground garnet, powdered walnut and other ground nut shells, grain hulls, aluminum oxide, plastic particles, and even tiny pieces of sponge are just a few of the other materials that have also been used as abrasive products [11]. Ice particles, or pelletized dry ice (carbon dioxide or CO2), are other mediums used as abrasive products, but we should recall that this methodology is too strong in most of the historic masonry [11].

62.2.4 Laser Cleaning Laser cleaning is another technique that is used sometimes by conservators to clean small areas of historic masonry. It can be quite effective for cleaning limited areas, but it is expensive and generally not practical for most historic building heritage [9, 12].

1796

F. Fratini et al.

62.2.5 Biocleaning Recently the possibility of using the microbial action to clean the surfaces of stone sculptures, buildings, and frescoes has been evaluated. In particular, the ability and potential of different microorganisms to remove undesired sulfates, nitrates, and organic matter have been studied [13]. Although it may seem contrary to common sense, masonry cleaning should be carried out starting from the bottom and proceeding to the top of the building always keeping wet all the surfaces below the area being cleaned. The rationale for this approach is based on the principle that dirty water or cleaning effluent, dripping from the above cleaning in progress, will leave streaks on a dirty surface but will not streak a clean surface as long as it is kept wet and rinsed frequently.

62.2.6 Environmental Considerations The potential effect of any method proposed for cleaning historic masonry should be evaluated carefully. Chemical cleaners and paint removers may damage trees, shrubs, grass, and plants. A plan must be provided for environmentally safe removal and disposal of the cleaning materials and the rinsing effluent before beginning the cleaning project.

62.2.7 Safety Considerations Possible health dangers of each method selected for the cleaning project must be considered before selecting a cleaning method in order to avoid harm to the cleaning operators, and the necessary precautions must be taken. Abrasive methods produce dust which can pose a serious health hazard, particularly if the abrasive product or the masonry contains silica.

62.3

Methods and Materials

Three bricks were selected on which control measurements were made before and after cleaning. For this purpose, a plastic mask was used to be able to reposition in the same measurement area. On each brick, three areas for color and water absorption measurements and three other areas for photographic survey and peeling measurements were identified (Fig. 62.1).

62.3.1 Color Measurements: Spectrophotometric Methods The color variations induced by the cleaning treatments on the selected three bricks are a purely aesthetic parameter that may be interesting to the final global perception

62

Cleaning of Masonry Surfaces of Cultural Interest

1797

Fig. 62.1 The façade of the building of via Montemagno (a) and detail of the area identified for the measurements (b). (Credit Fabio Fratini)

of the building facade. The Konica Minolta spectrophotometer CM 700d adopting the CIE L* a* b* method [14] was adopted for the calculation of the color parameters. According to this method, the color of a surface is described by three parameters: L* (0 to 100) represents the lightness, a* is related to the impulse of redgreen color, and b* is related to the impulse of yellow-blue color [15]. The total change of color is summarized by the parameter ΔE* calculated with the following equation: h i1=2 ΔE
¼ ðΔL
Þ2 þ ðΔa
Þ2 þ ðΔb
Þ2 where ΔL* ¼ (L*before-L*after); Δa* ¼ (a*before-a*after); Δb* ¼ (b*beforeb*after). The measurements were performed with diffuse illumination (D65 standard source) on an area of 8 mm in diameter, with specular component included and excluded. Three measurements for each selected area were carried out (nine measures for each brick). The measurements were made according to the following procedure: first in the position defined as zero, then rotating 45 degrees from this position, and rotating again by 45 degrees to return to the zero position (Fig 62.2a).

62.3.2 Superficial Cohesion Measurements: Peeling Test To assess the effect of cleaning on the surface cohesion of the bricks, peeling measures have been carried out. This test is used to evaluate the dusting of murals paintings in industrial environments, and it has been successfully applied to wall paintings and stone materials for the evaluation of the consolidating action of conservative treatments [16]. The 3 M, series 1280 Scotch tape was used, with

1798

F. Fratini et al.

Fig. 62.2 The test for the evaluation of cleaning treatment: color measurements (a), peeling test (b), and water absorption by contact sponge (c). (Credit Manuela Mattone)

52.5 N/cm2 of tensile strength and 3.3 N/cm2 of adhesion strength. The investigated peeling area is 15 cm2. Three measurements were taken, respectively, before and after cleaning (Fig 62.2b).

62.3.3 Water Absorption Measurements: Contact Sponge Method In order to measure the water absorption variations of the bricks induced by cleaning, the contact sponge method was chosen for its in situ ease of use. Following the UNI 11432 [17] standard, a 5.6 cm diameter sponge (24.6176 cm2 test area) was used, soaked in 5 ml of distilled water, and kept in contact with the surface for 30 s. Three measurements were made before and after cleaning (Fig 62.2c).

62.3.4 Aspect of the Surface: Photographic Survey In order to assess the effect of cleaning on the morphology of the surface, photos were taken with a camera Nikon COOLPIX P500 at close distance.

62.4

Results and Discussion

Tables 62.1, 62.2, and 62.3 and Fig. 62.3 show the results of the measurements made on the façade of the building in Via Montemagno/Buttigliera in Turin. First of all we have to say that the results point out a substantial unevenness of the bricks. In particular the color measurements, the water absorption, and the peeling tests are more homogeneous for bricks 1 and 2 than for brick 3. For this reason it is better to comment on the results for each brick individually rather than as an average of the three bricks, although for completeness the average is shown in the following graphs. Nevertheless, regarding the color measurements, a general trend followed by the three bricks can be emphasized (Fig. 62.4) with a decrease in brightness and an increase of a and b coordinates which is manifested by more red-yellow color.

62

Cleaning of Masonry Surfaces of Cultural Interest

1799

Table 62.1 Colorimetric measurementsa Test area Brick 1

Brick 2

Brick 3

Brick mean values a

Cleaning treatment Before After Δa Before After Δa Before After Δa Before After Δa

La 55.68  3.00 50.80  1.89 4.88  1.12 56.65  1.33 53.02  0.70 3.63  0.64 54.72  1.93 48.17  1.05 6.55  0.91 55.68  0.97 50.66  2.43 5.02  1.47

aa 4.18  0.10 10.33  0.33 6.15  0.43 3.70  0.20 7.10  0.27 3.40  0.09 3.84  0.34 8.11  1.26 4.27  1.15 3.91  0.25 8.51  1.65 4.61  1.41

ba 13.10  0.91 17.57  0.42 4.46  0.57 13.31  0.85 19.02  0.36 5.71  0.50 12.22  1.81 17.94  1.25 5.72  0.58 12.88  0.58 18.18  0.75 5.30  0.72

ΔEa 9.03  0.05

7.57  0.39

9.69  0.60

8.63  1.08

Average of nine measures

Table 62.2 Water absorption by contact sponge

Test area Brick 1

Test 1 2 3 Mean value Brick 2 1 2 3 Mean value Brick 3 1 2 3 Mean value Brick mean values

Amount of absorbed water Before After (g/cm2) (g/cm2) 0.607 0.785 0.596 0.782 0.626 0.812 0.609  0.015 0.793  0.016 0.578 0.785 0.590 0.788 0.596 0.791 0.588  0.009 0.788  0.003 0.569 0.776 0.580 0.779 0.588 0.785 0.579  0.009 0.780  0.005 0.592  0.016 0.787  0.007

Increase (%) 22.757 23.856 22.895 23.169  0.599 26.353 25.178 24.658 25.396  0.868 26.656 25.506 25.087 25.750  0.813 24.772  1.399

Therefore, from this point of view, the cleaning produced a renewal effect, making the brick of a color similar to that at the moment of the construction. As for the water absorption, there is a trend of the increasing amount of water absorbed after the cleaning (Fig. 62.5). This datum indicates that surely the cleaning had the effect of removing the incoherent dirt and crusts. As a matter of fact, the superficial deposits (dirt) provide a certain hydro-repellence due to filling the porosity but also to their nature (they are constituted also by hydrocarbons).

1800

F. Fratini et al.

Table 62.3 Peeling measurements Test area Brick 1

Test 1 2 3 Mean value Brick 2 1 2 3 Mean value Brick 3 1 2 3 Mean value Brick mean values

Amount of removed material Before (g/cm2) After (g/cm2) 0.066 0.067 0.066 0.067 0.066 0.066 0.066  0.001 0.067  0.001 0.066 0.066 0.066 0.067 0.067 0.067 0.066  0.001 0.067  0.001 0.061 0.067 0.060 0.067 0.060 0.067 0.060  0.001 0.067  0.001 0.064  0.003 0.067  0.001

Increase (%) 0.5 0.3 0.3 0.4  0.1 0.3 0.2 0.4 0.3  0.1 10.0 10.2 10.4 10.2  0.2 3.617  4.924

As for the amount of material removed with the test of the adhesive tape, it is possible to observe that the trend for bricks 1 and 2 is similar, while in the case of brick 3, there has been a slight increase. It is therefore possible that in this last case, the cleaning has weakened the surface or that a certain amount of superficial deposits was not removed by the final washing after cleaning (Fig. 62.6). All the macro images seem to indicate a general increase in the surface roughness of the bricks, a change that is in agreement with a greater ability to absorb water (Fig. 62.3).

62.5

Conclusions

The cleaning of building surfaces is an intervention often necessary for aesthetic reasons. In the case of buildings of historical and artistic interest, rules and protocols are applied in order to avoid their damage. In particular for these buildings, a preliminary study is carried out aimed to know the nature of the materials/type of dirt and to test the most suitable cleaning method. Instead, in the case of not protected buildings of historical interest, due to the absence of regulation, the cleaning methodology is decided just according to economic and temporal factors, and it is up to the owner of the building. Indeed, there are currently no municipal, regional, or state incentives that allow the owner to carry out basic diagnostics. Furthermore, the owner often lacks knowledge of the historical value of the building also due to the negligence of public institutions with regard to a disclosure in this sense. The choice of the operating

62

Cleaning of Masonry Surfaces of Cultural Interest

1801

Fig. 62.3 Particular of the surface roughness before (a, c, e) and after (b, d, f) the cleaning treatment for respectively brick 1, 2, and 3. (Credit Fabio Fratini)

mode can therefore not only have harmful effects on the building but also lead to a loss of the historicity/identity of the building itself. The aim of this contribution was to show how, with simple techniques (easy to execute, quickly, and with readily available materials), it is possible to obtain indications about the most suitable cleaning technique (effective, without causing damage). It is simply a matter of applying these diagnostic tools (peeling tests, water absorption, assessment of color change) with at least two different operating conditions or two different mechanical cleaning techniques in test areas before applying them over the entire façade. These techniques could be further simplified and become truly affordable for everyone.

1802

F. Fratini et al.

Fig. 62.4 Trend of changes in color due to the cleaning treatments (the values refer to the average of nine measurements for the brick area). (Credit Silvia Rescic)

Fig. 62.5 Trend of changes in water absorption due to the cleaning treatments (the values refer to the single measurements for brick test area). (Credit Silvia Rescic)

62

Cleaning of Masonry Surfaces of Cultural Interest

1803

Fig. 62.6 Trend of changes in the amount of material removed due to the cleaning treatments (the values refer to the single measurements for brick test area). (Credit Silvia Rescic)

References 1. Raccomandazione Normal 20/85, 1985. Conservazione dei materiali lapidei: Manutenzione ordinaria e straordinaria, CNR, Roma 2. Grimmer AE (1988) Keeping it clean: removing exterior dirt, paint, stains and graffiti from historic masonry buildings. Diane Pub Co., Washington, DC 3. Biscontin G, Zendri E, Bakolas A, Longega G, Driussi G, Moropoulou A (1995) Alcune considerazioni sulla pulitura delle superfici architettoniche, Proc. Scienza e Beni Culturali XI, Brixen, pp 625–631 4. Warren J (1999) Conservation of brick. Butterworth Heinemann, Oxford 5. Kyle C. Normandin & Deborah Slaton (Eds.) (2005) Cleaning techniques in conservation practice. J Archit Conserv 11(3):7–32 6. Revez MJ, Delgado J (2016) Rodrigues: incompatibility risk assessment procedure for cleaning of built heritage. J Cult Herit 18(219–228) 7. Ashurst N (1994) Cleaning historic building. Cleaning materials and processes, vol 2. Donhead, London 8. Casaletto MP, Ingo GM, Riccucci C, De Caro T, Bultrini G, Fragalà I, Leoni M (2008) Chemical cleaning of encrustations on archaeological ceramic artefacts found in different Italian sites. Appl Phys A 92:35–42 9. Ďoubal J (2014) Research into methods of cleaning silicate sandstones used for historical monuments. J Archit Conserv 20(2):123–138

1804

F. Fratini et al.

10. Iglesias-Campos MA (2014) Effects of mechanical cleaning by manual brushing and abrasive blasting on lime render coatings on Architectural Heritage. Mater Constr 64(316):e039 11. Pozo-Antonio JS, López L, Dionísio A, Rivas T (2018) A study on the suitability of mechanical soft-abrasive blasting methods to extract graffiti paints on ornamental stones. CoatingsTech 8: 335–345 12. Siano S, Agresti J, Cacciari I, Ciofini D, Mascalchi M, Osticioli I, Mencaglia A (2012) Laser cleaning in conservation of stone, metal, and painted artifacts: state of the art and new insights on the use of the Nd:YAG lasers. Appl Phys A 106:419–446 13. Cappitelli F (2016) Biocleaning of cultural heritage surfaces. Open Conf Proc J 7:65–69 14. CIELAB (1976) Commission Internationale de l’Eclairage 15. EN 15886:2010. Conservation of cultural property – test methods – colour measurement of surfaces 16. Drdácký M, Lesák J, Rescic S, Slížková Z, Tiano P, Valach J (2012) Standardization of peeling tests for assessing the cohesion and consolidation characteristics of historic stone surfaces. Mater Struct 45(4):505–520 17. UNI 11432:2011. Cultural Heritage Natural and artificial stone Determination of the water absorption by contact sponge

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient Pompeii

63

He´le`ne Dessales, Julien Cavero, and Agne`s Tricoche

Contents 63.1 63.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Architectural Repairs: A Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2.1 Repairs and Reconstructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.2.2 Reinforcement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.3 The OPUR Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.3.1 Design of the Database and Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.3.2 Organization of the OPUR Database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.4 Mapping the Repairs: A Multi-scale Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.4.1 A Field Data Acquisition Tool . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.4.2 Criteria for Recording and Mapping Repairs . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.4.3 Spatial and Quantitative Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1806 1807 1807 1808 1808 1808 1810 1815 1815 1817 1818 1819 1820

Abstract

The restoration of a building entails specific techniques and professional skills, revealing the logic and logistics of reconstruction. Focusing on the repairs made following a major earthquake allows us to explore such a building system, which is implemented at times of crisis and emergency. From 2015 to 2019, the research program RECAP (“Reconstruire après un séisme. Expériences antiques et innoH. Dessales (*) Ecole normale supérieure, AOROC, PSL Research University, Paris, France e-mail: [email protected] J. Cavero CNRS, LGP, Paris, France e-mail: [email protected] A. Tricoche CNRS, AOROC, PSL Research University, Paris, France e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_63

1805

1806

H. Dessales et al.

vations et à Pompéi”; Rebuilding after an earthquake. Ancient experiments and innovations in Pompeii) was carried out to shed new light on the reconstruction conditions in Pompeii and building techniques in the wake of an earthquake. This contribution sets out the methodology of the project, covering three main steps: the definition of an architectural repair; the presentation of the analytic criteria of the repairs through the database OPUR (Outils pour unités de réparation; Tools for repair Units); and mapping post-seismic interventions in Pompeii. Promoting a multidisciplinary approach to built heritage, this method may prove helpful to archaeologists, architects, and engineers for characterizing repairs in areas prone to seismic risk. Keywords

Reconstruction · Repair · Building techniques · Earthquake · Risk · Database · GIS · Pompeii

63.1

Introduction

Among the various possible natural disasters, earthquakes are a frequent cause of degradation to built heritage, depending on its level of vulnerability. For several decades, the reconstruction of damaged buildings has been subject to precautionary measures and para-seismic standards, as detailed in national recommendations and published in various handbooks [1, 2]. For ancient periods, however, be it in Antiquity or in the Middle Ages, there seems to have been no treaty – at least none that has been preserved – providing the technical solutions to be adopted after a seismic event. In the absence of normative sources, therefore, only the archaeological observation of standing remains can shed light on the knowledge held by builders. There are many possible solutions, ranging from makeshift patching up to technical innovations. These empirical experiences depend not only on economic and cultural factors but also on the frequency of seismic incidents, which could generate a major sense of vulnerability in a specific place [3]. Three significant factors seem to determine this culture of reconstruction: the level of qualification of the builders [4, 5]; the availability of manpower and materials; and the possibilities of financing the building processes, managing demolition work, and dealing with emergencies [6]. The RECAP program, funded by the Agence Nationale de la Recherche [7], addresses these key questions by tracing Roman post-earthquake reconstructions. To develop this approach, the project focused on a paradigmatic site, Pompeii. Pompeii is a unique laboratory: this small town in Campania suffered several earthquakes which followed on from one another at short intervals in 20 years prior to the eruption of Vesuvius in 79 AD. While the first, dated 62/63 AD and probably the most violent [8], is well known thanks to written sources, the subsequent episodes have only been identified through archaeological data [9, 10]. The exemplary nature of this case study should be emphasized, since it demonstrates “resilience” in progress and indeed at the very moment of being interrupted, since the

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

1807

reconstruction work necessitated by the successive seismic episodes was brutally interrupted by the final eruption. An exceptional “snapshot” of the process can thus be examined. When retracing the history of Roman construction, scant interest has been paid to this topic [11]. Thanks to the exceptional state of conservation and sheer extent of the remains, the site of Pompeii is particularly suited to a systematic investigation. This is true both on the small scale, with a focus on specific construction sites as case studies, and on the large scale, taking a complete overview of the management of building processes in the ancient town. On the small scale, we chose to focus on two complementary case studies: firstly, priority urban services revealed by the water network as well as through the extra-urban aqueduct [12] and the 15 public urban water towers and secondly, the private building strategies attested to by a vast aristocratic residence – namely, the Villa of Diomedes [13]. On the large scale, we applied a systematic approach to an extended area of Pompeii, visualizing all the post-seismic repairs. We selected the Regio VII with its 10 blocks (insulae n 1, 2, 3, 4, 9, 10, 11, 12, 13, and 14), to the east of the forum (5 ha, around 15% of the discovered site), as a representative area that comprises public and religious buildings, houses, shops, and industries. In this way, we were able to compare the reconstruction conditions in the various functional categories of the city [14]. In the field, the investigation is based on a pluridisciplinary approach: the archaeology of standing remains; earth sciences; earthquake engineering; and computer vision. This brief contribution will focus on the protocol developed in the framework of the RECAP project for the characterization of post-earthquake reconstructions in the Regio VII. Three main, complementary methods will be presented: the definition of an architectural repair, in the context of Pompeii; the analysis of the repairs using a specific database; and the cartography of the post-seismic interventions.

63.2

Architectural Repairs: A Definition

In recent years, research on the concept of repairs has intensified [15]. A repair is a special moment in the “biography of an object,” whatever its nature, be it an item of furniture or a building [16, 17]. It makes it possible to restore the functionality of an object that has been damaged through wear, a defect, or an accident. Such a transformation generates sociability, knowledge, and skills, revealing the overall organization of a production system. Characterizing the architectural repairs made following a major earthquake allows us to grasp this production system relating to the reconstruction of an urban environment, both in a state of crisis and emergency, especially when key infrastructures are affected.

63.2.1 Repairs and Reconstructions Repairing a building involves replacing a damaged structure, so as to more or less reproduce its original configuration [18]. An initial distinction should be introduced

1808

H. Dessales et al.

here, according to the scale of the repair. If the repair of the structure is wholesale, meaning that the entire structure is rebuilt, it is defined as a reconstruction. If the repair of the structure if partial, it could be of various types: – The restoration of wall corners – The refection of opening jambs or lintels – The filling of cracks on walls, pillars, columns, arches, or vaults

63.2.2 Reinforcement Another type of post-earthquake intervention needs to be singled out here: reinforcements. Such interventions are not made on the damaged structure itself but rather involve the addition of an external element to consolidate or reinforce. These might include: – – – –

The filling of openings Wall redoublings Buttresses Arch buttresses

In the case of Pompeii, we therefore produced an inventory of the ancient repairs observed on the standing remains. We excluded anomalies that have not been repaired, such as cracks or partial collapses, for instance, those restored in contemporary times. Indeed, they may relate to other seismic episodes than the 62/63 AD earthquake, such as the 79 AD eruption of Vesuvius or more contemporary episodes, such as the earthquake of Irpinia in 1980, which had a significant impact on Pompeii [19]. Only repairs carried out in the ancient period were therefore taken into account. In order to carry out this inventory in the field, two complementary tools were developed: firstly, a database designed to record the different types of repair and reinforcement and secondly, a mapping interface to map these interventions in the ancient city of Pompeii.

63.3

The OPUR Database

Through the analysis of the repairs, the OPUR database has the distinctive feature of focusing on the indirect effects of earthquakes. By this way, it allows to go back to direct effects, through the characterization of the related damage.

63.3.1 Design of the Database and Availability The OPUR database (“Outil Pour Unités de Réparation”; Tool for Repair Units) was designed at the same time as the observations were carried out in the Regio VII of Pompeii and then immediately tested in the field. The database partially reproduces

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

1809

the structural organization of the OPUS database (“Outil Pour Unités Stratigraphiques”; Tool for Constructed Stratigraphic Units). The latter was developed to record the stratigraphic units and employed during the campaigns conducted on the Villa of Diomedes in Pompeii between 2012 and 2016 [20]. However, we ruled out a systematic stratigraphic approach, since it was impossible to identify all the stratigraphic units of the buildings in the Regio VII. Instead, a more specific approach was developed using the OPUR database in order to record the various repair units identified in a building. The OPUR database was designed using the software FileMaker and deposited on a remote server made available via the French project known as TGIR (“Très Grande Infrastructure de Recherche”; Very Large Research Infrastructure) Huma-Num, which is dedicated to digital technologies in the field of the humanities and social sciences [21]. OPUR was first tested during a field mission in July 2016, which focused on the Building of Eumachia (VII 9, 1). It was then used during a 3-week campaign in July 2017, in order to record all the visible repairs in the ten selected insulae of the Regio VII. In the field, OPUR users, divided into three teams, worked on PC tablets with the help of locally installed clone versions and an adapted data display: at the end of each investigation, the data were checked and merged on the shared version. In total, 255 repairs were recorded in the Regio VII. More than 1500 photographs produced during the field campaigns were added to the database, which makes it possible to display them according to each repair unit consulted. After the field campaign, the full version of the database was made available and could be edited by the whole team, who are able to consult and update the database with restricted access, in a collaborative process. In its finalized version, the OPUR database describes each repair according to a system of indexing and document analysis developed in collaboration with the structural engineers associated with the RECAP project (Fig. 63.1). We propose a version that can be applied to any archaeological site, in order to allow for the rapid processing of information collected on the various architectural repairs. This database, produced in three languages (French, English, Italian), is now publicly available to everyone, in free access, from the site RECAP [7]. We hope that this tool will be useful to archaeologists, architects, and engineers engaged in the analysis of repairs and that it will promote a multidisciplinary reading of built heritage. Below, we have presented the main choices adopted for characterizing the metadata of each repair unit in OPUR. Ready-to-use lists of values were defined. They can be consulted either through the RECAP database or by consulting the paper record sheet (Fig. 63.2). The latter is particularly suitable for the analysis of elevations (walls, buttresses, columns, and pillars) and openings (doors, windows, and niches), where the most damage and repairs can be observed in Pompeii (67% and 30% of the repairs in the Regio VII, respectively). This paper record makes it possible to test and adapt the method in the field, whether or not the information is intended to be recorded in the database. The database itself offers several additional functionalities, such as image management, and the possibility of duplicating a record or automatically constituting a catalog in PDF format, after selecting the data and images to be included in the final presentation, where appropriate.

1810

H. Dessales et al.

Fig. 63.1 Diagram of the organization of the OPUR database (A. Tricoche, and A. Picandet)

63.3.2 Organization of the OPUR Database RECAP consists of four distinct parts: identifying the repair; the damage; the nature of the repair; and chronological relations (See Fig. 63.1).

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

OPUR Wall Doorway

REPAIR UNITS Buttress Window

Column Niche

AUTHOR(S) & DATE

ID. NUMBER

Pillar

IDENTITY Site, building:

In the case of a wall Position 1 :

Geo coordinates (x, y, z): Type of intervention:

Partial

Static function 2 :

Full In the case of a partial intervention

Description:

Orientation of the repaired part: Height of the repaired part 3 :

DAMAGE Nature of the disorder 4 : Precisions on the type(s) of damage (wall, doorway, window, niche) 5 :

Angle (°):

Ancient origin 6 :

Probability of the event 7 :

REPAIR In the case of a partial intervention

Dimensions (cm): Length:

Width:

Thickness:

Diameter:

Height:

Technique:

Nature of the repair 8 :

Preparatory intervention:

Type of technique 9 :

Presence

Materials, precisions 10, 11 :

Precisions:

RELATIONS Links to the Atlas of Roman Building Techniques database, ACoR (or other database):

Links to other OPUR repairs:

The exponents refer to the corresponding value lists.

Fig. 63.2 (continued)

COMMENTS, NOTES

1811

1812

H. Dessales et al.

OPUR [1] POSITION (walls) Facade wall Interior wall

[2] STATIC FUNCTION (walls) Load-bearing wall Non load-bearing wall

[3] HEIGHT OF THE REPAIRED PART Upper part Middle part Lower part

[4] NATURE OF THE DISORDER Inclination Deformation Fracture Crack Supposed collapse Settling

[5] TYPE OF DAMAGE If angle of 2 walls Corner overturning If wall alone Vertical crack Diagonal crack in single / double direction Rocking Disgregation of masonry Flexural cracks Sliding shear failure Partial overturning of wall with rotations axis at floor level Global overturning of the façade Global or partial overturning of single leaf with rotations axis at the floor level or at the base of facade Partial complex overturning Global complex overturning of the façade Vertical arch mechanism Local overturning of wall due to tie beam hammering Horizontal arch mechanism If walls leaning against another Irregularities in plan and elevation, related to the addition of other stacked buildings If arch Crack in the impost Crack in the keystone Crack in the haunch

[6] ANCIENT ORIGIN Earthquake Vertical movement of the ground

VALUE LISTS

[7] PROBABILITY OF THE EVENT Very strong Strong Low Very low

[8] NATURE OF THE REPAIR If elevation Corner Cracks filling Creation of buttress Reconstruction Wall redoubling Arch buttress If opening Frame Jamb Lintel Support Threshold Blocking

[9] TECHNIQUE OF REPAIR Block Masonry Filling-in Metallic tie

[10] MATERIALS OF REPAIR Floor fragments Plaster fragments Mortar Stone Terra cotta If terra cotta Amphora Brick Tile Other ceramic If mortar Lime Sand Broken terra cotta Earth

[11] TYPE OF METALLIC TIE Cramp Dowel Iron Lead

Fig. 63.2 Paper record sheet of the OPUR database (A. Tricoche)

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

1813

Identifying the Repair The repair is first determined by its identifier number and location (both the type of building element where the repair was observed and its location on the site). In the case of Pompeii, the distinction was made by the Regio, Insula, and property number (civico). In the more general version available for download, however, which is intended for other sites, we replaced these values with the geographical coordinates (x, y, z). In addition, any other reference or nomenclature can be indicated to link the repair with a parallel indexing system or a preexisting typology using an ID number. Thus, for the study of the Regio VII in Pompeii, for example, we have indicated the link with the database ACoR [22], developed to record Roman building techniques, which will be made available online in 2020. This international project led to the creation of a controlled, hierarchized vocabulary for defining and characterizing building techniques in the Roman world, in the form of a new trilingual thesaurus (French, English, and Italian) [23]. Damage Multidisciplinary in nature, the RECAP project involved a team of archaeologists and structural engineers, the aim being to determine the nature of the damage and its potential link with an earthquake. To specify the nature of the damage, various specific cases are envisaged: for example, in the case of damage affecting a wall, where observations are most common, it is necessary to distinguish between an angle between two walls, a wall alone, or walls leaning against one another, each of which is connected to a list of values; in the specific case of an arch, the location of the crack in the arch can be specified. The nature of the damage is indicated according to various disorders: inclination; deformation; fracture; crack; supposed collapse; and settling. The different types of damage are listed with reference to the codification proposed in the “Linee guide per la valutazione e riduzione del rischio sismico del patrimonio culturale” [24, 25], and a schematic illustration is provided for each type by Giuseppina De Martino and Francesca Autiero, under the direction of Andrea Prota (Università degli Studi di Napoli Federico II, Dipartimento di Strutture per l’Ingegneria e l’Architettura). Following this observation, the ancient origin of the event (earthquake or vertical movement of the ground) is identified and its probability evaluated [26]. Repair In the case of a partial intervention, a description of the nature of the repair can be provided. This description can be either free or associated with a list of values (especially for elevations and openings). For example, the values for an elevation are as follows: corner; crack filling; creation of buttress; reconstruction; wall redoubling; and arch buttress. The dimensions of the repair may also be indicated (according to the individual case: length, width, height, thickness, diameter), as well

1814

H. Dessales et al.

as the type of technique used (blocks, masonry, metallic ties). For each of these categories, a list of identified materials used for the repair is proposed, which can be enriched according to the specificities of each archaeological site. Finally, an indication and description of a preparatory intervention observed for the repair can be given.

Chronological Relations Two types of chronological relations are established. First, an external link can be made to the ACoR database or any other reference, indicating, by means of a code, the types of construction techniques defined, as indicated supra. Thus, a repair can be posterior, anterior, or contemporary to one or more types of construction techniques (Fig. 63.3). By establishing these relationships, it is possible to develop a statistical approach to identify the most damaged types of techniques, as well as the types of techniques most commonly used in the case of repairs. For example, in the case of Pompeii, terra cotta masonry appears to have been favored in wall repairs (40%).

Fig. 63.3 Example of an OPUR unit identification (Casa delle Nozze d’Ercole, VII 9, 47, Atrium, part of the east wall): repair consecutive to a corner overturning, OPUR unit no. 8 (cf. Fig. 3) (H. Dessales, photograph T. Crognier, 2017, published with the permission of the Ministero per i Beni e le Attività Culturali e per il Turismo – Parco archeologico di Pompei)

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

1815

An internal link is then proposed within the framework of the OPUR database, as a repair may be posterior, anterior, or contemporary to one or more other repairs. In the specific case of Pompeii, the identification of 11 repairs successive to previous repairs, all related to a seismic episode, tends to prove the succession of several earthquakes in the final years of the city [9, 10, 13]. The metadata chosen to characterize and describe the repair units can be used, to some extent, to automate analyses and expand upon the results obtained. There are various possibilities, according to the needs. Thus, for our research in the Regio VII of Pompeii, we were able to establish a typology of partial repairs, which was generated computationally by crossing the data still available on the database (the type of construction element combined with the nature of the repair). This typology consists of four types, which feed in automatically (repair of wall corners; repair of opening jambs; cracks filling of cracks; reconstruction of elements). The same autotypological work was carried out to characterize the type of reinforcement associated with the repairs observed (filling of openings; wall redoubling; buttress; arch buttress).

63.4

Mapping the Repairs: A Multi-scale Approach

The systematic mapping of the various repair units identified in the field and recorded in the OPUR database was associated with a geographic information system (GIS) (ArcGIS v.10). The use of a GIS is required here to undertake spatial and quantitative analyses of post-seismic interventions and seismic impacts at the scale considered. To achieve this result, several objectives had to be met. The aim was, from a technical point of view, to have an effective registration interface that could be implemented directly in the field and, from a methodological point of view, to define the elements that needed to be surveyed. In fact, in order to produce this survey of all the post-seismic intervention units throughout the Regio VII, it was necessary to determine and define the various criteria for mapping the interventions, in addition to the information recorded in the OPUR database.

63.4.1 A Field Data Acquisition Tool Given the size of the urban area selected for the survey – 10 insulae covering more than 5 ha and corresponding to more than 300 properties – it was necessary to employ a field data acquisition tool that was both mobile and adapted to the architectural scale. It also needed to be quick to implement in the field. The solution chosen was a collector for ArcGIS, a free geographic data collection application for mobile devices developed by ESRI. The collector principle is based on editing pre-configured geographical entities (geometry and attribute table) on a background map loaded into the mobile device (Fig. 63.4). If the latter is equipped with a photographic sensor, photographs can be attached directly to the entities surveyed.

1816

H. Dessales et al.

Fig. 63.4 Field recording interface based on a 1/50 plan (collector for ArcGIS) (J. Cavero)

The data collected are saved in SQLite format and synchronized – or postsynchronized in an offline mode – with the GIS project. Due to the urban nature of the area surveyed, and even more so for specific surveys produced inside some buildings, it was not possible to use positioning using a Global Navigation Satellite System. To overcome this constraint, we decided to produce a 1/50 plan adapted to a multi-scale operation, coupled with an aerial photograph of the area and other information making it easier to determine locations in the field’, such as civico numbers associated with a geocoding system. This plan was established on the basis of the surveys carried out using the documentation kindly provided by the Archaeological Park of Pompeii, within the framework of the “Nuova cartografia informatizzata georiferita” initiated as part of the “Grande Progetto Pompei (Piano della Conoscenza)” [27]. The geo-referenced vector plans were imported into the GIS) and combined into several feature classes, respecting the original categories. After conducting a topological simplification to remove overlapping areas, the feature classes were classified by floor. Finally, a range of visibility was associated with each one in order to obtain a dynamic basemap that offers several levels of detail depending on the display scale. The accuracy of this plan was first tested and validated on the Building of Eumachia (VII 9, 1) and then more broadly on the public buildings of the forum in 2016. The basemap was then published as a tile map service and made available on a PC tablet. The repairs identified could then be mapped in the field in parallel with the registration in the database. Collector for ArcGIS allows several operators to work simultaneously and for the data to be synchronized afterward. For reasons of

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

1817

efficiency, though, a single operator was employed to control the cartographic survey and to ensure systematic photographic coverage of each element.

63.4.2 Criteria for Recording and Mapping Repairs The mapping of the post-earthquake repair interventions of the Regio VII was carried out during a single 3-week field mission in 2017 using this basemap, which offers 1/50 scale accuracy. However, prior to the first surveys, it was necessary to define the elements to be collected so that the data entered in the GIS could be used to answer the key questions underpinning the project: to specify the organization of post-seismic reconstruction projects on an urban scale; to define the vulnerability of the structures; and to determine the intensity of the seismic impacts. Criteria for recording spatial data were thus defined and adapted to ensure that the GIS was truly complementary to the information recorded in the OPUR database. Four categories of elements that are directly relevant to postseismic repairs and constructions were thus selected: actual repairs; associated reconstructions; associated constructions; and wholesale construction sites corresponding to the last phases of the city. Each repair corresponds to a unique record in the database, unlike the other three categories, which have only been mapped. – Repairs are defined as intervention units following a damage whose seismic origin is considered as probable. Each repair is characterized by its technical specificity and its position within a construction element. A stratigraphic logic was adopted, such that two units present in the same elevation but having no point of contact were treated as two repairs. – Reconstructions associated with a repair are those elements that belong to the same building process as a specific repair unit and that correspond to an identical reconstruction of previous elements. Reconstructions related to space redevelopment were therefore not included. – The constructions associated with a repair are those elements that belong to the same building process as a specific repair unit but for which it is not possible to specify whether they are an identical reconstruction of previous elements. – Post-seismic building processes consist of elements whose construction techniques and materials are characteristic of the last phase of the city and feature new reconstructions, without any integration of previous structures. This is the case, for example, with some of the public buildings in the forum (VII 8), which have been rebuilt with a terracotta masonry of cut tiles [4, 6, 28]. These distinctions need to be made in order to develop the most accurate quantitative spatial approach possible. In addition, depending on its nature, each element is part of a type of geographical, and therefore geometric, expression. Constructions (including last phase construction sites) and reconstructions are recorded in a polygonal shape. The same applies to repairs whose surface area is sufficient to be mapped at 1/50. These elements can then be characterized by their length, orientation, and volume. On the other hand, crack filling, block repairs, or

1818

H. Dessales et al.

interventions on columns or pillars are recorded in a punctual shape. For each element, the ID generated by the database was reported in order to connect the two recording systems. Finally, a specific reference to the geographical data was set up to link constructions and reconstructions associated with a specific repair.

63.4.3 Spatial and Quantitative Analysis The unique spatial database set up as part of the RECAP project allows for a large number of observations and analyses (Fig. 63.5). The cross-referencing analysis of information on the types of damaged techniques and the types of techniques used for repairs are all results that can be spatially translated. The various post-seismic construction sites can then be characterized and included in the urban history of the last phase of the city. A spatial analysis of the types of elevation damaged (facades, dividing walls, isolated elements) or the estimated volumes of materials used can also be produced. The analysis of the post-seismic repair interventions recorded is carried out using a multi-scalar approach in order to clearly define their organization and propose a mapping of the impacts of seismic activity on the site: at the scale of the building unit, in order to understand the vulnerability of each construction and the repair techniques used; at the scale of the block (insula), in order to define the link between the different constructions, as well, and perhaps above all, as to consider the differences in topographical level that exist in the city due to the presence of underground spaces in the various buildings; lastly, at the scale of the entire city, to compare the function of buildings and the distribution of the different construction sites following the earthquake.

Fig. 63.5 3D view of post-seismic interventions identified in the Regio VII (repairs in red, associated constructions in orange) (J. Cavero)

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

63.5

1819

Conclusion

Combined, the perspectives provided by the different methods adopted shed new light on the adaptation strategies adopted by Roman society in the wake of an earthquake, by defining the typology of repair techniques used. As such, the variability of the techniques employed becomes apparent, depending on the nature of the damage and the category of the building (public or private) in question. By using this approach, it is possible to evaluate the capacity of ancient societies to recover after a natural catastrophe, i.e., their “resilience.” It also allows us to appraise the economic impact of earthquakes – a topic that is currently at the center of debates, with Pompeii having constituted a case study ever since the pioneering article by J. Andreau [29]. Indeed, urban reconstructions constitute an excellent barometer for observing economic conditions. Mapping and recording the repairs reveals a whole new aspect of Pompeii, establishing a dynamic picture of a town under construction and involving infrastructures, dwellings, the urban fabric, and the landscape. Finally, the methods applied in Pompeii to record and map the reconstruction of the urban system allow for an approach that is both qualitative and quantitative. On the one hand, we have tried to qualify the various repairs and reconstructions observed, proposing a typology of the post-seismic interventions. On the other hand, these results have allowed us to quantify the level of damage observed in the various building and to determine macroseismic intensity, according to the different scales available (EMS 98 for example) [30, 31]. For the 62/63 AD earthquake in Pompeii, preliminary studies indicated an intensity of IX in the MCS scale [19, 32]. Its magnitude was evaluated as 5 [33, 34]. Thanks to the large scale of our study and the quantification of the repaired structures and destructions, we can propose a level of urban destruction of 20% (still to be confirmed), with a degree IX MCS; on the EMS scale, the intensity would be of IX, considering a damage level on grade D5 (vulnerability of the structures: level A/B). Developed on the unique laboratory that is Pompeii, this new protocol – designed to identify repairs and reconstructions– could be extended to other ancient sites in areas prone to seismic risk, whichever historical period is being considered. By highlighting the transformations and adaptations a building undergoes over time, it makes it possible to examine both the preventive techniques implemented to reduce the vulnerability of a building and improvised restorations that may be devoid of any consideration of risk. Such an approach could promote a new reading of built heritage, by underscoring the relation between natural risk and construction. Acknowledgements We wish to extend our warmest thanks to Laura Pecchioli for her invitation and help with producing this chapter. Our gratitude also goes to the entire RECAP “Regio VII” team: Francesca Autiero, Guilhem Chapelin, Marina Covolan, Arnaud Coutelas, Thomas Crognier, Giuseppina De Martino, Marco Di Ludovico, Christophe Loiseau, Florence Monier, Arnaud Montabert, and Andrea Prota. Without the kind support of the team at the Archaeological Park of Pompeii, this program might never have seen the light of day. We would like to thank its director, Massimo Osanna, excavations director Grete Stefani, architecture officer Annamaria Mauro, archaeology officers Laura De Esposito, Bruno De Nigris, Giuseppe Scarpati, and scientific assistant Vincenzo Sabini. Finally, we thank Arnaud Picandet for the illustration of this article as well as Victoria Weavil for her revision of the English text.

1820

H. Dessales et al.

References 1. Davidovici V (2016) Conception, construction parasismique. Traité d’application de l’Eurocode, vol 8. Eyrolles, Paris 2. Tesfamariam S, Goda K (2013) Handbook of seismic risk analysis and management of civil infrastructure system. Woodhead, Cambridge 3. Ferrigni F, Helly B, Victor LM, Pierotti P, Rideaud A, Teves Costa P (eds) (2005) Ancient buildings and earthquakes. The local seismic culture approach: principles, methods, potentialities. Edipuglia, Bari 4. Dessales H (2011) Les savoir-faire des maçons romains, entre connaissance technique et disponibilité des matériaux. La connaissance des roches et son application par les structores. Le cas pompéien ». In: Monteix N, Tran N (eds) Les savoirs professionnels des hommes de métier romain. Centre Jean Bérard, Naples, pp 41–63 5. Dessales H (2012) Not built in a day: awareness of vulnerability and construction techniques in Roman times. In: Carvais R, Guillerme A, Nègre V, Sakarovitch J (eds) Nuts & bolts of construction history. Culture, technology and society, vol 3. Picard, Paris, pp 471–477 6. Dessales H (2015) La produzione laterizia a Pompei. Adeguamento di un materiale e organizzazione dei cantieri urbani. In: Bukowiecki E, Volpe R, Wulf-Rheidt U (eds) Il laterizio nei cantieri imperiali. Roma e il Mediterraneo. Archeologia dell’architettura 20. Insegna del Giglio, Sesto Fiorentino, pp 81–89 7. RECAP Project. REConstruire Après un tremblement de terre: Expériences antiques et innovations à Pompéi. Rebuilding after an earthquake. Ancient experiments and innovations in Pompeii), Agence Nationale de la Recherche (ANR ANR-14-CE31-0005, 2015–2019). http:// recap.huma-num.fr 8. Savino E (2009) Nerone, Pompei e il terremoto del 63 d.C. In: Marino AS, Mérola D (eds) Interventi imperiali in campo economico e sociale. Edipuglia, Bari, pp 225–244 9. Fröhlich T, Jacobelli L (eds) (1995) Archäologie und Seismologie. La zona vesuviana dal 62 al 79 D.C. Problemi archeologici e sismologici. Biering und Brinkmann, Munich 10. Varone A (2005) Convivere con i terremoti. La travagliata ricostruzione di Pompei dopo il terremoto del 62 d.C. alla luce delle nuove scoperte. In: Mols STAM, Moormann EM (eds) Omni pede stare. Saggi architettonici e circumvesuviani in memoriam Jos de Waele. “L’Erma” di Bretschneider, Naples, pp 315–324 11. Giuliani CF (2011) Provvedimenti antisismici nell’antichità. Journal of Ancient Topography 21: 25–52 12. Filocamo F, Carlut J, Dessales H, Borensztajn S (2018) Roman builders facing the risk of disaster. Coupling archaeological and geochimical analyses of a section of the « Aqua Augusta » (the Roman Serino aqueduct, Southern Italy). Archaeometry 60(3):915–932 13. Dessales H (ed) (2020) The Villa of Diomedes. The making of a Roman villa in Pompeii. Hermann, Paris 14. Dessales H (ed) (forthcoming) Ricostruire dopo un terremoto. Riparazioni antiche a Pompei. Centre Jean Bérard, Naples 15. Ambers J, Higgit C, Harrison L, Saunder D (eds) (2009) Holding it all together: ancient and modern approaches to joining, repair and consolidation. Archetype, London 16. Appadurai A (ed) (1996) The social life of things. Commodities in cultural perspective. Cambridge University Press, Cambridge 17. Kopytoff I (1996) The cultural biography of things. Commoditization as process. In: Appadurai A (ed) The social life of things. Commodities in cultural perspective. Cambridge University Press, Cambridge, pp 64–91 18. Roulon-Doko P (2007) Les mots de la réparation. In: Speranza G (ed) Objets blessés. La réparation en Afrique. Musée du quai Branly, Paris, pp 19–23 19. Adam JP (1989) Osservazioni tecniche sugli effetti del terremoto in Pompei del 62 d.C. In: Guidoboni E (ed) I terremoti prima del Mille in Italia e nell’area mediterranea. SGA, Bologna, pp 168–171

63

Post-Earthquake Reconstruction: Mapping and Recording Repairs in Ancient. . .

1821

20. Tricoche A (2020) Database management system. In: The Villa of Diomedes. The making of a Roman villa in Pompeii. Hermann, Paris, pp 121–132 21. Dessales H, Tricoche A (2018) Un database per studiare le riparazioni post-sismiche. Archeologia dell’architettura 23:19–24 22. “Atlas des techniques de la Construction Romaine”; Atlas of Roman Building. https://acor. huma-num.fr/ 23. Camporeale S, Dessales H, Tricoche A (eds) (forthcoming) Atlante delle tecniche della costruzione romana. Manuale. Qasar, Rome 24. Ministero per i Beni e le Attività Culturali (ed) (2010) Linee guida per la valutazione e riduzione del rischio sismico del patrimonio culturale. Gangemi, Rome 25. Papa S, Di Pasquale G (2011) Manuale per la compilazione della scheda per il rilievo del danno ai beni culturali. Chiese-Modello A-DC. Presidenza del Consiglio dei Ministri, Dipartimenti della Protezione Civile 26. Arrighetti A (2015) L’archeosismologia in architettura. Per un manuale. Firenze University Press, Florence 27. Morichi R (2018) Il disegno di una città antica: le scelte grafiche per una gestione dinamica delle informazioni. In: Morichi R, Paone R, Sampaolo F (eds) Pompei. Nuova cartografia informatizzata georiferita. Arbor Sapientiae, Rome, pp 139–158 28. Maiuri A (1942) L’ultima fase edilizia di Pompei. Istituto di studi romani, Spoleto 29. Andreau J (1973) Histoire des séismes et histoire économique; le tremblement de terre de Pompéi (62 ap. J.-C.). Annales (ESC) 28:369–395 30. Grünthal G, Levret A (2001) L’échelle Macrosismique Européenne. European Macroseismic Scale 1998. Conseil de l’Europe, Luxembourg 31. Guidoboni E, Ebel J (2009) Earthquakes and Tsunamis in the past. A guide to techniques in historical seismology. Cambridge University Press, Cambridge 32. Guidoboni E, Ferrari G, Mariotti D, Comastri A, Tarabusi G, Sgattoni G, Valensise G (2007) CFTI4Med. Catalogue of strong earthquakes in Italy (461 BC – 1997) and Mediterranean Area (760 BC – 1500). INGVA-SGA, Bologna. http://storing.ingv.it/cfti/cfti5/quake.php?50039IT 33. Cubellis E, Luongo G, Marturano A (2007) Seismic hazard assessment at Mt. Vesuvius: maximum expected magnitude. J. Volcanol. Geotherm. Res 118:339–351 34. Marturano A, Rinaldis V (1998) Seismicity before the 79 AD Vesuvius eruption. In: Albore Livadie C, Ortolani F (eds) Il sistema uomo-ambiente tra passato e presente. Centro Universitario Europeo per I Beni Culturali, Ravello, pp 237–246

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana Tomb

64

Ivan Roselli, Lucina Giacopini, Alessandro Colucci, Vincenzo Fioriti, Marcello Melis, Tiziana Pasciuto, Gerardo De Canio, Angelo Tatı`, Francesca Boitani, and Laura D’Erme

Contents 64.1 64.2 64.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Study Area and Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applied Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64.3.1 3D Laser Scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64.3.2 Ambient Vibration Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64.3.3 Magnified Motion Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64.3.4 Hypercolorimetric Multispectral Imaging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64.4 Experimental Application to the Campana Tomb . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1824 1825 1830 1830 1831 1833 1835 1838 1842 1848

Abstract

A multidisciplinary approach was used for the study of an important Etruscan tomb located in Veii, Italy. It included the application of several noncontact and/or noninvasive technologies and methods with the aim of investigating two different I. Roselli (*) · A. Colucci · V. Fioriti · G. De Canio · A. Tatì ENEA, Rome, Italy e-mail: [email protected]; [email protected]; vincenzo.fi[email protected]; [email protected]; [email protected] L. Giacopini Around Culture Srl, Rome, Italy e-mail: [email protected] M. Melis · T. Pasciuto Profilocolore Srl, Rome, Italy e-mail: marcello.melis@profilocolore.it; tiziana.pasciuto@profilocolore.it F. Boitani Former SBAEM, Rome, Italy L. D’Erme National Etruscan Museum of Villa Giulia, Rome, Italy e-mail: [email protected] © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_64

1823

1824

I. Roselli et al.

aspects of the site’s conservation: the structural stability and the preservation of the wall paintings. After documenting the 3D geometry of the site by laser scanning, some ambient vibration recordings were analyzed, and the innovative technique of magnified motion analysis (MMA) was applied to evaluate the structural conditions of the tomb. In particular, these analyses have been conducted to identify the structural behavior of the walls and ceiling supporting the paintings. Frequency response function (FRF) and horizontal-vertical spectral ratios (HVSR) were calculated to estimate the fundamental frequencies. MMA were applied for Image processing, revealing small motions of points of walls and ceiling that may indicate exfoliation and detachment phenomena. The wall paintings were investigated through hypercolorimetric multispectral imaging (HMI). Integrating the results of the two technologies, a valuable contribution to the documentation and the diagnosis of the paintings was provided. In particular, the innovative HMI technique was able to extract and compare the spectral reflectance of pigments against a specialized spectral database of pigments’ spectral signatures. Besides, it was very effective in enhancing the readability of very ruined paintings, almost invisible to the naked eyes.

64.1

Introduction

In the context of conservation and prevention of cultural heritage, the present chapter emphasizes the importance of a multidisciplinary approach in order to give the most complete and valuable contribution to site managers to deal with their archaeological sites. In particular, the interaction between site managers and scientists that made available their different expertise to measure/analyze different aspects of the site through several noninvasive techniques and digital technologies stimulates mutual cooperation and ideas to improve focusing of scientific research, on one side, while, on the other side, and more importantly, leads to a better knowledge of the overall site. More in detail, the activity presented in this chapter was conducted in the framework of interdisciplinary collaborations between different domain experts – archaeologists, researchers, physicists, chemists, engineers – within the archaeological project at the Veii’s Necropolis site (Parco Naturale Regionale di Veio, Rome, Italy). This cooperation was established in the scenario of the COBRA regional project promoted by ENEA and developed in partnership with Around Culture Srl), a SME based in Rome and specialized in cultural heritage, with the aim to support archaeological documentation, analysis, and dissemination regarding the Veii’s archaeological remains and provide an inspiring environment for promoting partnership in this field. The project involved two of the three monumental Veii’s sepulchers that preserve evidence of Etruscan painting: the Tomba dei Leoni Ruggenti (or Roaring Lions tomb) and that preserve the oldest evidence of Etruscan painting dating back to the seventh century BC, and the Tomba Campana (or Campana tomb) with the aim of applying several diagnostic techniques employed in the cultural heritage. All these

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1825

techniques are characterized by noninvasiveness, response speed, sensitivity, and selectivity for many classes of materials. The synergistic use of multiple techniques makes it possible to obtain further information about the main constituents of the paint layer, pigments, binders, consolidating materials, protective materials, dyes, biodegradation, etc. The following chapter describes the multidisciplinary approach applied to the case of the Tomba Campana, which presents some structural damages, and the paintings on the walls are not currently visible due to the fading of the original colors. In fact, the current knowledge of the paintings is based more on descriptions and sketches made by artists and researchers who had visited the tombs from the nineteenth to the twentieth century. The activities focused on the application of several scientific technologies and methods, some of which are innovative. After a 3D laser scanning for geometry investigation, the structural issues of the tomb were evaluated. In particular, the analysis of ambient vibration and the magnified motion analysis (MMA) were carried out. Also, advanced technologies for the documentation and diagnosis of the painted walls, such as the hypercolorimetric multispectral imaging (HMI), were applied. The above technologies are described and their application is discussed. Ambient vibrations and MMA have been conducted to identify the structural behavior of the walls and ceiling supporting the paintings. Frequency response function (FRF) and horizontal-vertical spectral ratios (HVSR) were calculated to estimate the fundamental frequencies. MMA were applied for image processing, revealing small motions of points of walls and ceiling that may indicate exfoliation and detachment phenomena. The HMI analysis has been conducted for reading of the main radiometric and colorimetric properties at pixel level of painted surfaces. This technique is able to extract and compare the spectral reflectance of pigments against a specialized spectral database of pigments’ spectral signatures. Besides, it is very effective in enhancing the readability of very ruined paintings, almost invisible to the naked eyes. Due to the artistic and archaeological importance of the site, specific and detailed studies were also required to scientifically validate the status of the tomb. Furthermore, 2D digitization monitoring analyses and surveys were necessary to record and understand the state of conservation of the tomb, but at the same time obtain a useful cataloging and training tool. The collaboration between ENEA and Around Culture within the COBRA allowed a multidisciplinary study in the famous paintings of the Tomba Campana which produced valuable information useful not only for artistic hypotheses but also for archaeological and conservation purposes.

64.2

Study Area and Site

The Veio’s site is an outstanding testimonial of the Etruscan culture, in which with the beginning of the Iron Age, around 920–900 BC, the urban formation process is clearly evident on the great plateau of the historic city, parallel to the development of vast burial grounds located on the surrounding hills. In many sectors of the hills, the necropolises

1826

I. Roselli et al.

have been identified and partly excavated: thousands of sepulchers, of the well and pit type, with their funeral kit have been brought to light, dating back at the end of 10th until the end of 8th. It is at the end of this century that the first underground chamber tombs appear, many of which have remains of painted walls. The monumental sepulchers currently visible are few in number compared to the thousands of Veii tombs dating back to the protohistoric era. Exceptions are some hypogea that preserve the oldest evidence of Etruscan painting dating back to the seventh century BC, such as the Tomba delle Anatre (or Tomb of the Ducks) of the necropolis of the Riserva del Bagno, the Tomba Campana in the necropolis of Monte Michele, the Tomba dei Leoni Ruggenti in the necropolis of Grotta Gramiccia, the oldest painted hypogeum dating back to the early decades of the seventh century, which exhibits a frieze of animals rendered in a “contour line” in the most ancient painting technique mentioned in the sources [1–5]. The hypogea are open to public visits in rotation, due to issues of conservation [6]. In the tombs natural lighting is very limited or totally absent, while artificial illumination is often inappropriate for a complete and detailed observation of the painted walls. In the following, the multidisciplinary diagnostic tests carried out in the Tomba Campana, one of the most emblematic and controversial sepulchers of the archaeological site of the Necropolis of Monte Michele, are described (Fig. 64.1).

Fig. 64.1 Map of the Veii plateau (from [7]): the arrow shows the position of the Campana Tomb (a). Plan (bottom) and Frontal (top) views. Frontal view of the inner entrance wall after drawings by Francesco Caracciolo del 1825 ca (b)

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1827

The tomb, which was dated 620–610 BC, was officially discovered in 1843 by Giovanni Pietro (or Giampietro) Campana, an art collector expert of ancient Rome, and is widely known for its context and its extraordinarily painted walls [8, 9]. The tomb is characterized by an articulated plan, with two chambers aligned on the same longitudinal axis and two cells on the sides of the dromos, cut into the volcanic stone (Fig. 64.2). The pictorial repertoire of the tomb is characterized by a surprising palette of colors in a delicate state of preservation, essentially due to the microclimate of the hypogeum, which caused damage to the structure until the collapse of some parts of the vault and the loss of cohesion of the painted layers. For the reasons set out above, access to the tomb is currently limited to the two underground burial chambers, which are those that still preserve the murals, unlike the other two side chambers. Two stone statues depicting two lions are placed at the end of the dromos as guardians of the front door of the burial chambers (Fig. 64.3a) and from the beginning an internal wall with large polygonal stones reinforced the entrance wall. The discovery of the tomb with depositions and grave goods was the subject of numerous publications made a few months after the discovery and subsequently based on the progress of the studies [10–16]. The figures and the descriptions included in such publications still remain fundamental, as they constitute the most important and detailed documentary source of the monument. The latest work of reinterpretation of the data published in the nineteenth and twentieth century was the one recently conducted by Eline Verburg, which aimed at reconsidering the controversial issue of the discovery and the modern history of the tomb [17]. In particular, the authors’ novel intuitions along the preceding investigation lines [16–22] confirm and add further elements to the open debate about the hypothesis that the discovery of the Tomba Campana is an organized event rather than an accidental finding. As for the analysis of the painted surfaces, the interest was mainly focused on the paintings in the first chamber of the hypogeum. Its decorative ensemble is now substantially lost. But some visible traces remain, even if almost

a

b Second chamber

NORTH

First chamber

Lateral chamber

Entrance

Dromos

Fig. 64.2 Campana tomb: frontal view from the dromos (a) and plan view (b)

Lateral chamber

1828

I. Roselli et al.

Fig. 64.3 The entrance to the Campana Tomb in a repertoire photo from [21] (a) and the rear wall of the first chamber as it appears today (b)

unrecognizable to the naked eye (Fig. 64.3b). Their colors are red, black, yellow, and white/gray (mostly faded), which underline what remains of the orientalizing painting cycle that is considered the most remarkable of its times. As known, the rear wall of the first chamber presented the most ancient Etruscan megalography. It essentially comprised four framed squares, two of them on each side of the rectangular door in the rear wall of the first inner chamber, working as a passage toward the second and more inner chamber (Fig. 64.4a): it alludes to the journey of the dead towards the Kingdom of Bliss. On both sides of the door, which is framed with geometric decorations mainly made up of vividly colored triangles painted directly on the tuff wall, the wall paintings presented, below a decorative band with lotus flowers, a human character riding a horse with an elongated body, while the bottom square displays with fantastic animals. The man is represented alone on horseback in the left wall while on the right is surrounded by his cortege in a context of “departure for the hunt,” to the exaltation of his rank of “prince,” among real, fantastic beasts and a lush intertwining of plant elements: the deceased is followed by a dog and has a small jaguar crouched on the horse’s back and is accompanied by a bridegroom holding the bridle and a bearer with a double-edged ax, the emblem of the deceased prince. In the second chamber, a series of shields painted on two lines highlight the high rank of the deceased. The current knowledge of the tomb is substantially based on descriptions and sketches by artists and researchers that visited the site since the nineteenth century. These documents provide information on the state of conservation and the limited painting remnants at the time of each visit. Briefly, the most relevant ones can be summarized as follows. The first known images of the site are the ones by L. Canina [10], who reproduced accurate and detailed drawings (Fig. 64.4a). Also to be mentioned is the illustration by L. Piroli [11], providing an overall image of the

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1829

Fig. 64.4 Frescoes of the rear wall in the first chamber, according to drawing by L. Canina [10] (a), and overall view of the chamber’s contents in place (drawing by L. Piroli [11]) (b)

first chamber with the tomb’s contents still in place (Fig. 64.4b). Even more interesting are the drawings published by Micali [12]. They display the upper right part of the wall paintings, maybe supplied by Campana himself, presenting some slight differences from the Canina’s drawings, such as the yellow points against a field of red to hatch the anterior part of the horse, absent in Canina’s work. A later documentation is the watercolor by Alessandro Morani, dated back to 1897 (Fig. 64.5a), which recently appeared in an exhibition organized by the Swedish Institute. It was probably intended to provide a more realistic representation of the paintings, but rather distorted by the prevailing gray color of the same nuance of the composition’s background. A drawing (lucido) in 1:1 scale, made by architect Giorgio Wenter Marini in 1915, when, just graduated in Germany, he went to Rome for his internship before entering the Marcello Piacentini’s study, is reproduced in a photo that was found by Laura D’Erme in the archives of the Museo di Villa Giulia, Rome: it shows the rear wall of the first chamber (Fig. 64.5b). As the quality of Wenter Marini’s work is not very high, uncertainties in the tract and misunderstandings can be confused with gaps in the pictorial film. The latest photographic documentations that testify a still quite good state of conservation of these paintings date back to the 1950s and 1960s. Unfortunately, these are black-and-white pictures, so no information about the color can be retrieved. To get a more recent picture, we must refer to the study conducted by F. Banti at the beginning of the 1970s of the last century that gives a detailed representation of

1830

I. Roselli et al.

Fig. 64.5 Watercolor by A. Morani, 1897 (a), and a drawing in 1:1 scale by architect G. Wenter Marini, 1915 (b)

the pictorial scheme elaborated on the segments of the preserved painting that remains the current reference for his illustration [18–20]. Similar to the tomb’s contents, the hypothesis that the paintings are not completely genuine, but arbitrarily adjusted and revived at the time of their discovery, has been proposed [21, 22]. Such hypothesis is not to be totally dispelled, even if in contrast with the results of analytic investigations conducted in the 1970s, when the colors were already highly altered and degraded [19].

64.3

Applied Methods

64.3.1 3D Laser Scanning The 3D laser scanning is a contactless, nondestructive technique capable of reconstructing the geometry of objects through a cloud of points provided with the xyz coordinates. The instrument has a laser probe emitting an impulse of light in a given direction. The emitted light hits the object’s surface and, after being reflected by the object, partially back to the instrument, is acquired, and the time needed in the process is recorded (time of flight). Such time is utilized to calculate the distance of the object from the instrument, once that the speed of light is known, in that direction. The system is also equipped with controlled servomotors that rotate the instrument’s mirrors, allowing a scanning of the volume around the acquisition station. The use of this technique is by now very consolidated to document accurately the geometry of cultural heritage assets, even in the application to large buildings and entire archaeological sites. If one acquisition station is not enough to cover the entire investigated site with a single scan, then several stations can be planned, and the partial point clouds are merged together through algorithms based on rototranslation operations of each scan. To this aim, homologous points acquired in two or more scans (recognizable points normally materialized with reflecting targets) are needed.

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1831

The instrument is generally equipped also with a high-resolution camera that is calibrated in accordance to the internal parameters of the scanning system, so as to provide a colored realistic rendering of the point cloud. In order to use the acquired data for volume and surface calculations, the point clouds are transformed into threedimensional representations of the object that can be post-processed with various software packages suitable for specific applications. In the application to cultural heritage conservation, this technique is a useful tool for monitoring changes in geometrical parameters that can indicate instability and/or degradation of sites (e.g., the verification of the verticality of walls, changes in overall volumes that might indicate partial collapses, changes in the distance between different parts due to settlements or structural deformations, etc.).

64.3.2 Ambient Vibration Analysis Ambient vibrations are originated from a series of sources that are present in the general environment. In contemporary urban areas, they are mainly associated to transportation infrastructures (e.g., road traffic, metro systems, urban railroads, tramways, etc.). Other relevant vibration sources in the urban environment are related to construction activities and crowded events. In sparsely populated or low-density areas, such as in the countryside, vibrations from anthropic activities are obviously more limited. Generally speaking, there are also a number of natural vibrations that are due to wind and seismic activities. Through the ground the ambient vibrations are transmitted from the source to the structure foundations and from foundations to the upper structure. They represent a major factor of fatigue in the materials, which can also lead to dangerous damage to the structures, especially in the case of constructions characterized by high vulnerability. In some circumstances, this might be the case of historic buildings and monuments. Consequently, in the last several decades, conservationists, scientists and engineers increased their interest in studying these phenomena. Contemporarily, lawmakers or regulatory agencies also released guidelines and regulations that indicate limits of vibration intensity that human activities and transportation systems are allowed to release to the environment, with a special focus on sources placed close to very vulnerable structures. On the other hand, vibrations may cause nonstructural damage to buildings and their content, such as in the case of cultural heritage assets represented by very delicate objects, wall paintings, etc. In such cases, even small damages, e.g., mere cosmetic damages, may be a major problem for both monetary and immaterial values. This is certainly the case of archaeologic sites and their content. In the literature, several experimental and theoretical works recently aimed at studying the effect of ambient vibrations and related phenomena on buildings can be found [23–28]. Several authors studied the effects that external vibration sources may induce on historical buildings. According to Rainer [29], a distinction must be considered between short-term and long-term effects, taking into account that the latter ones could gradually produce damages even if the vibration

1832

I. Roselli et al.

intensities are very low [30]. This observation is of crucial importance for archaeologic sites where ruins and buildings may not be structurally sound and materials degradation processes may already be undergoing due to outdoor exposure to atmospheric agents. In order to face the impact of urban vibrations on the structural health of buildings, several regulations and standards were adopted in different countries that decided to limit the vibration intensities at significant reception points of the structures. Indications are given about the methods for the measurement procedure (e.g., proper type of instrumentation, location of measurement points, data processing and analysis, etc.). An internationally accepted parameter used to assess the intensity of vibration is the peak particle velocity (PPV), defined as the maximum value of vector velocity recorded in triaxial acquisitions. Also widely accepted, and substantially analogous parameter, is the peak component particle velocity (PCPV), defined as the maximum value of velocity recorded in each triaxial direction. Other aspects that international standards take into account are the dominant excitation frequencies, the types of building under examination (the strictest limits are allowed for historic structures), and the duration of the vibrations (short-term and long-term). Even though national regulations present different limits of acceptability, especially the ones concerning ancient buildings, as the structural characteristics of this kind of buildings are very different from one country to another, as a general indication, most of current international standards and recommendations suggest that long-term vibration intensity at historic buildings should be limited to PPV or PCPV values below 2–3 mm/s [31–33]. Another important use of ambient vibration acquisition is for obtaining a dynamic characterization of structures. In fact, ambient vibration excites the structures, making them oscillate according to their own natural modes, which constitute a dynamic property of the structure itself, since they essentially depend on the geometry and on the mechanical properties of the materials. Therefore, the modal parameters (e.g., the natural frequencies) are widely used for the structural health monitoring of historic buildings [34]. The vibration data acquired at several points of the studied structure may be processed with a variety modal analysis techniques in order to extract such modal parameters that can provide indications on the state of the overall structure. Also, different points of the structure may give different responses, providing indications that parts of the monument have different structural behavior locally, which may suggest investigating the possibility of structural damage and/or mechanical disconnections between that part and the remaining structure. A common formulation of the frequency response function (FRF) with the transmissibility function H was used to process and analyze the vibration data in order to extract the modal frequencies [35]. Ambient vibration acquisitions are also used to characterize the response of the ground. In particular, the horizontal-to-vertical spectral ratio (HVSR) method is based on the ratio between the amplitude of the Fourier spectra of horizontal H and vertical V components recorded on a structure (also known as H/V technique). This method was initially introduced by Nakamura [36] to determine the

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1833

predominant period of ground motion signals to evaluate site amplification characteristics during earthquakes. Subsequently, it was successfully explored to estimate also the modal frequencies of buildings, on the basis of the simple observation that resonance at lower structural modes (generally the most relevant ones for the overall dynamic behavior of the building) causes vibration amplification in horizontal directions, while the system remains much stiffer and substantially unamplified in the vertical direction [37]. It is an empirical technique able to provide rough estimates of natural frequencies on the basis of the hypothesis that vertical vibration travels through the structure without significant amplification in comparison to the horizontal components, which are much more amplified. In practice, this assumption remains substantially valid for bending modes in transversal directions when soilstructure interaction is negligible, as usually happens in the case of ambient vibrations.

64.3.3 Magnified Motion Analysis To the aim of improving the monitoring of vibrations induced to historic structures and monuments for structural health survey and early damaging detection, a new methodology of machine vision processing is available, namely the motion magnification (MM) technique [38, 39]. Motion magnification (MM) acts like a microscope for tiny motions in digital videos: small displacements, not visible to the naked eye become clear and evident after the magnification processing. The methodology does not rely on optics but on algorithms capable to amplify only the tiny changes in the video frames, while the large ones remain. It is not a new discipline in the field of the analysis of mechanical structures and buildings, but till a few years ago, it was not a viable one. Since any pixel can provide an intensity variation time series, each pixel can be considered a contactless “virtual sensor.” Thus, a very large number of such sensors are made available for an ex post analysis, meaning that it is not necessary to decide in advance the sensor placements, because every point of the surface recorded on the video will provide a signal. Of course these signals may be processed by the conventional quantitative analysis techniques such as the analysis in the frequency domain. However, the extension of the MM to large structures is not straightforward: dimensions, illumination, distance from the recording device, and vibration disturbances induced on the video cameras are all issues to be taken into consideration carefully [40]. The basic MM version looks at intensity variations of each pixel and amplifies them, revealing small motions which are linearly related to intensity changes through a first-order Taylor series for small pixel motions. The major limitation is the linear approach entailed in Taylor’s expansion that set an upper limitation to the amplification. In practice, to remain in the linearity bound, we need slowly changing images and small amplifications. Moreover, physical limitations, such as illumination, shadows, camera unwanted vibrations, poor pixel resolution, low frame rate, presence of large motion, and distance from the object, decrease severely the quality of the motion magnification and should be taken into account in

1834

I. Roselli et al.

order to achieve high-quality results. Actually, during many laboratory tests reported in the literature, MM was effective in amplifying subtle motions in videos, making distinguishable the tiny displacements of structures exposed to mechanical disturbances and facilitating the evaluation of the structure stability within a reasonable time. The first step of the MM analysis is to record a video of the structure including the critical points we are interested in, taking good care to avoid large motions such as people passing by in front of the camera, swinging cables, or non-fixed mechanical parts. The presence of large motions is the most significant source of noise for the MM, requiring the experimenter to isolate part of the image, although usually this is not a feasible option. At the moment the issue is still open. Indoor illumination is also a key aspect: shadows should be avoided as much as possible. These time series contain the information about the displacements of the physical points related to the pixels that allow a complete frequency domain analysis. Of course, it would be too cumbersome to analyze all the virtual sensors, and moreover not all the surface of the structure generates useful information; therefore, we identify a small area of the surface with a high signal-to-noise ratio. In the outdoor environments, things are more complicated; nevertheless, MM has proved to be a reliable and affordable tool also in these circumstances. All these results show that the recent developments in digital vision technologies are very promising. In short, the advantages provided by MM are many: a large number of “virtual sensors” available, no wires, reduced amount of data storage, no physical contact, simplicity, low costs, and predictive capabilities. Unfortunately, noise is still a pervasive obstacle to MM analysis, especially when the recording device is a low-quality video camera, but in the near future, it is likely that research will provide a hardware implementation to strongly reduce noise and be capable of real-time functioning. Therefore, we may look at the motion magnification as one of the most important improvements in the field of contactless vibration monitoring. Recently, a number of experiments conducted on simple geometries like rods and other small objects, as well as on bridges, showed the reliability of this methodology compared to accelerometers and lasers, but in the case of analysis in the open environment such as for the Etruscan tombs and generally for the cultural heritage, it is much more complicated. The historic monument protection is a primary task, and the first step surely is to arrange a suitable monitoring. Besides the usual wear due to age, traffic, and weather, we have also to face earthquakes as a major threat [41]. Consequently, researchers are trying continuously to ameliorate the monitoring systems, hopefully in the sense of enlarging and extending their predictive capabilities: to this end, we propose the use of motion magnification. The analysis of image sequences in the field of civil engineering is not new. For many years attempts to produce qualitative (visual) and even quantitative analysis using high-quality videos of large structures have been conducted, but with poor results. This was because of the resolution in terms of pixels, of the noise, of the camera frame rate, of the computer time, and finally of the lack of appropriate algorithms able to deal with the extremely small motions related to building displacements. These and other limitations have restricted the applications of digital vision methodologies to just a few

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1835

cases. Nevertheless, recently important advances have been obtained by Freeman and collaborators of the Massachusetts Institute of Technology (MIT). From a practical point of view, it is necessary to record a digital video of high quality, avoiding or reducing mechanical or optical disturbances as much as possible [42]. Clearly, this is not a simple task in the outdoor environment, especially for monuments as the Etruscan tombs, lacking the direct light of the sun. Anyway, the procedure is simple in principle: inside the tomb is placed a source of artificial light, then some videos are recorded by a 50 fps camera, pixel resolution 720
1280. Using a 50 fps video camera, the maximum frequency allowed is 25 Hz: above this threshold will be introduced spurious frequencies because of the aliasing; furthermore, since in our elaborations the frequency resolution is 0.1 Hz, the video duration must be at least 10 s. The basic methodology is to take advantage of the large number of pixels contained in an image. Theoretically, we could have 921,600 “virtual sensors,” meaning that each pixel has a time history of intensity variation (color or gray scale), from the first frame to the last one. These time series contain the information about the displacements of the physical points related to the pixels. It would be too cumbersome to analyze all the virtual sensors, and moreover not all the surface of the structure generates useful information; therefore, we identify a small area of the surface with a high signal-to-noise ratio (SNR).

64.3.4 Hypercolorimetric Multispectral Imaging The hypercolorimetric multispectral imaging (HMI) is an innovative technology that enables a dramatic improvement in the use of cameras and filters in many different application fields, allowing to change the use of cameras from equipment to capture pictures to highly reliable imaging radiometric instruments [43, 44]. By now the main application fields that Profilocolore investigated and developed are cultural heritage, crime scene investigation, precision farming, and medicine, but other industrial fields will benefit from this technology in the near future.

Description of Technology and Achievable Results The HMI system was developed starting from some basic research about how would appear an ideal colorimetry if not derived from the human eyes physiology (Fig. 64.6a). Moving from this question and defining a three channels sampling of the visible spectrum to minimize the metamerism (optimal spectra sampling), we arrived at the definition of the ideal colorimetry, largely shared and accepted by the international scientific community, based on equienergy and evenly spaced color matching functions (Fig. 64.6b). Having defined the ideal colorimetry in the range of 400–700 nm visible light, we extended the same concept to the 300–1000 nm spectrum corresponding to the sensitivity of silicon sensor camera. We called them the seven hypercolorimetric matching functions that allow the best spectra sampling in this range of wavelengths (Fig. 64.7).

1836

I. Roselli et al. Color Matching Functions CIE

2 1.8 Z

1.6

Y

X

1.4 1.2 1 0.8 0.6 0.4 0.2 0 400

450

500

550 Wavelength (nm)

600

650

700

Ideal Sensitivity Curves 1 0.9

Relative Sensitivity

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 400

(c) 2013 Profilocolore Sri

500

450

550

600

650

700

Wavelength (nm)

Fig. 64.6 Color matching functions CIE (a) and ideal sensitivity curves (b)

Hypercolor Matching Function (by Profilocolore)

1

Relative Sensitivity

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

300

400

500

600 700 Wavelength (nm)

Fig. 64.7 Hypercolorimetric matching functions

800

900

1000

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1837

Applying strong and robust mathematical processing, mostly based on artificial intelligence, and starting from only two RGB shoots taken with custom optical filters, Profilocolore has been able to achieve an overall behavior of the system that matches the 98% of these ideal sensitivity functions, leading to a very high precision sampling of any spectra in this range. Besides this, using the same two shots and applying here strong calibration mathematics, we achieve CIELAB (D65 illuminant, 10 observer) colorimetric data for each pixel, with an average error bounded within an error of dE ¼ 2. The whole analysis process is very fast and can be described in three stages, described in the following subsections.

Image Acquisition Once identified the subject to study and analyze, the perimeter of the area under study is marked using an arbitrary number of spectrally measured white patches. These white patches will be used by the calibration software to correct any spectral and intensity heterogeneity of the light distribution. The source of light is represented by a number (usually two) portable flashes, modified to freely emit their full range of spectra, UV (about 300 nm) to near infrared (about 1300 nm). This guarantees the right energy covering the whole spectrum of sensitivity of the system. Inside the perimeter delimited by the white patches it is positioned also as a color checker. This is a checker made of colored patches coming from the set of the 1950 colors of the NCS (the Swedish Natural Colour System) catalog. Each patch has been carefully spectrally measured in our laboratories using bench instruments that are traceable to international metrology institutes. The whole scene, including also the reference elements, is acquired with two photographic shots, using two optical filters, whose spectral transmittance is custom and optimized to achieve the most performing calibration results from the software. If the UV-induced fluorescence is also of interest, a third shot is performed applying dedicated UV filters in front of the flashes that restrict the passing band between 300 and 380 nm, while the camera is equipped with visible only bandpass filter, in the range of 400 to 700 nm. This guarantees a safe gap of 380–400 nm to avoid any possible pollution of the final image by excitatory flashes. Image Calibration Starting from the two shots with the two custom filters, the calibration software SpectraPick is able to generate 7 bands of spectral reflectance centered at 350, 450, 550, 650,750, 850, and 950 nm, with a radiometric absolute precision above 95%, and an Adobe RGB color image with an average precision of dE ¼ 2. In other words, each individual pixel of the scene has its own spectral signature in the range of 300–1000 nm and its own colorimetry. Image Analysis Thanks to the calibration process, we can know accurate spectral reflectance and colorimetry of every detail of the scene. This allows, using PickViewer, our multispectral image analysis software tool to get out of the data a considerable amount of

1838

I. Roselli et al.

information and even radiometric relationships among points across the whole image. Some of the analyses include the reading of spectral signature at every pixel position or the average of a small square window centered on the pixel; any kind of mapping starting from a pixel and including every pixel of the image that shows a similarity in the spectrum or colorimetry with the chosen one, within a variable tolerance threshold; analysis of combined bands using the principal component analysis; segmentation and clustering of the image based on any set of original features (bands, colors), and even integration of colorimetric and multispectral images with data coming from other sensors (XRF, RX, Raman, THz, etc.).

64.4

Experimental Application to the Campana Tomb

In the present section, the experimental application of the multidisciplinary approach through several noninvasive techniques is illustrated. After visual inspections and interaction between researchers and site managers who provided past documentation and data on the site, besides their knowledge about the history and problems of the tomb, the most appropriate experimental techniques were chosen, and related campaigns for each investigation could be optimally focused and planned. The tomb presents structural problems evidenced by partial collapses of the ceiling of the main inner chamber and of some walls of the lateral chambers that occurred in the past. As a first step, in order to provide an updated documentation of the historical site, the 3D geometry was acquired through laser scanning technique. The instrumentation used comprised a Riegl Z360 equipped with a Nikon D100 digital camera. The nominal angular resolution is 0.0025 horizontal and 0.002 vertical, while range accuracy was +/ 6 mm. The first step of the scanning procedure was to decide the acquisition stations in order to cover the investigated volume in its entirety. To this purpose, three stations were chosen: one outside of the tomb (approximately in front of the tomb entrance) and two inside (one in the middle of each inner chamber). Two of the acquisition stations are illustrated in Fig. 64.8. In Fig. 64.9 details of the point clouds rendering of the scans at the entrance and at the first inner chamber stations are illustrated. The structural investigation of the tomb was mainly conducted analyzing the ambient vibration data and the video footages processed with the MM technique. The ambient vibration measurements were carried out with five digital seismographs equipped with triaxial velocimeters (Fig. 64.10). The acquisitions were performed in two configurations, named config I and config II. The positions of the instruments (numbered as 100, 101, 103, 104, and 105) are illustrated in Fig. 64.11. In config I the measurement points were positioned at ground level and on the tomb roof. One instrument (105) was located in one of the lateral chambers overlooking the dromos, one at the entrance door (103), and one in the middle of the main inner chamber (104). Two instruments (100 and 101) were placed on the roof in correspondence with both stone beds in the inner chamber. Subsequently, the instruments were positioned in config II, with three instruments (101, 104, and 105) on the stone ledge of the second inner chamber and one instrument on each stone bed of the first chamber (100 and 103). Vibration data

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1839

Fig. 64.8 3D laser scanning acquisition stations: one outside the tomb (a) and one inside (b) each inner chamber

were acquired for a duration of about 20 min at a sampling frequency of 200 Hz. The measured ambient vibration resulted in very low intensity, as expected in such a very little anthropic area and quite far from the main roads, with velocity peaks lower than 105 mm/s (Fig. 64.12). The vibration analysis in terms of the frequency response of the measurement points is depicted in Fig. 64.13, where the horizontal-to-vertical responses resulted substantially similar but not exactly the same, confirming that different portions of the structure have a slightly different dynamic behavior, which suggests possible mechanical disconnections due to deep cracks in the stone. The application of the MM technique was focused on the localization of the zones of the ceiling of the Campana tomb that show possible effects of natural vibrations on the structural stability, in order to protect visitors from the risk of stones falling. After a preliminary visual inspection a portion of the ceiling that was evidenced to be more problematic (i.e., presenting large cracks associated with voids apparently due to partial falls of stone) was investigated by MM (see Figs. 64.14 and 64.15). In Fig. 64.16 the investigated portion of the ceiling is viewed before and after the MM processing. The two frames are practically identical at first sight, as the vibration motion is extremely limited. Nonetheless, through a detailed quantitative analysis of the MM video some indications could be obtained. In particular, within the investigated portion of the ceiling, a region-of-interest (ROI) was selected, as done for the red box, and during the MM elaboration, this ROI will be the object of the stability analysis. A particular video processing will show how much this area of the ceiling is sensible to vibrations and consequently how much the fall from the ceiling is probable. Inside the ROI we identify the most important fracture lines by means of the pixel time series, represented in Fig. 64.17. Of course, major displacements

1840

I. Roselli et al.

Fig. 64.9 Details of point clouds rendering of the entrance (a) and of the first inner chamber (b) scans

indicate the more dangerous points. The automatic procedure is shown in Fig. 64.17: from the time series mean and variance are calculated for each pixel time series. Their maximum coincides with the largest displacements; therefore, the pixels are identified and highlighted in the red area of Fig. 64.17 and in detail in Fig. 64.18. The points of the ceiling corresponding to these pixels follow the most dangerous fracture lines and hence deserve more care. As for the analysis of the painted surfaces, the interest was mainly focused on the frescoes in the first chamber of the hypogeum (Fig. 64.4a). Its decorative ensemble is now substantially lost. But some visible traces remain. Their colors are red, black, yellow, and white/gray mostly faded), which underlines what remains of the orientalinspired painting cycle that is considered the most remarkable of its times. Besides the structural stability of the site, in the present study, the fresco paintings of the inner chambers were investigated. In particular, special focus was put on the readability of the wall paintings on the right of the inner door.

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1841

Fig. 64.10 Ambient vibration measurement station with a triaxial seismograph

CONFIG. I (ground level)

CONFIG. I (roof) NORTH

CONFIG. II (ground level) 105

104

101

X Y

101

100

104

ENTRANCE

103

100

ENTRANCE

103 ENTRANCE 105

Fig. 64.11 Schematic plan views of the Campana tomb with the ambient vibration acquisitions points: the positions of triaxial seismographs (instruments 100, 101, 103, 104, and 105) are indicated in both configurations I and II

The applied tests were meant to verify whether the vivid chromaticity characterizing the realistic and fantastic beasts, as well as the plant elements, was obtained through the use of only three pigments of red, yellow, and black colors. The other colors were supposedly obtained with red points against a field of black, yellow points against a field of red, and red points against a field of yellow. Moreover, it was to be ascertained whether the ancient artists made use of some kind of white color for the background, as some critics hypothesized even though this is not visible to the naked eye by now. Different points of the painted walls were analyzed with the HMI technique. In particular, the analyzed points are located in correspondence to the scene in the mid band paintings on the right of the door to the second inner chamber. The HMI

1842

I. Roselli et al.

Fig. 64.12 Velocity time history of instrument 101 in configuration I acquisition

instrumentation was positioned viewing toward the limited areas of the fragment where the traces of pigment still appear visibly colored. The aim was to identify and separate the different graphic elements that make up the analyzed portion of the painted cycle, by assigning the proper spectral signature to pigments and substrate. The actual physical conditions of the wall surface don’t allow to easily apply every step of the HMI acquisition procedure, and specifically it was not possible the positioning of white patches to correct the distribution of light on the surface, so special attention has been paid to evenly distribute the light on the surface. The positioning of color checker required a holder that occupied a portion of the image (Fig. 64.19a). Nevertheless, the calibration process obtained very accurate radiometric and colorimetric images that allowed to distinguish even the less relevant changes of values related to the presence or absence of pigments on the wall. A number of enhancing techniques have been applied to the purpose of distinguishing residual fragments of red and yellow pigments. Building a map of all red and yellow fragments yielded to a dramatic improvement of the readability of the shapes present in the original paintings, as is evident in Fig. 64.19b, where the HMI processed image is compared to the corresponding color photo showing what is visible to the naked eye (Fig. 64.19a). The above results are very encouraging, and it would be worth to extend the HMI investigation to all the walls where latent paintings are possibly present.

64.5

Discussion

The multidisciplinary approach utilized in the study of the archaeological site of the Campana tomb provided several interesting results that constituted a valuable contribution to the remarkable effort of site managers to achieve a deeper

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1843

Fig. 64.13 Frequency response analysis of the ambient vibration acquisitions of configuration I (a) and II (b)

1844

I. Roselli et al.

Fig. 64.14 Schematic plan view of the inner chambers of the Campana tomb. The area of the ceiling that was selected to be processed by the MM is indicated in the red box

NORTH

celling portion investigoted by MM

ENTRANCE

Fig. 64.15 The dotted-line red box indicates the area selected to be processed by the MM technique

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1845

Fig. 64.16 A frame of the red dotted box of Fig. 64.7 of the original video (a) and the same frame of the magnified video (b)

Fig. 64.17 Identification of features with high displacements. Up, ROI pixels time series on a time-displacement graph, highlighted the larger displacements; bottom left, mean of the pixel timeseries; bottom right, variance of the time-series; bottom center, a MM video frame, the red the area is indicated by the maximum mean and variance of the time series

understanding of the analyzed monument both in terms of historic and material knowledge. Such improved knowledge of the site is essential also to plan conservation and preservation interventions. According to the analysis of the ambient vibrations, the absence of anthropic activities comprising relevant vibration sources in the surrounding of the site constitutes a favorable situation, at least in comparison to the archeologic site located

1846 Fig. 64.18 False colors of the bottom-center image in Fig. 64.17. The major cracks in the stones are evidenced in the red box

Fig. 64.19 Position of color checker (a) and HMI processed image (b)

I. Roselli et al.

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1847

within densely populated urban areas. Nonetheless, raining water infiltrations and the invasiveness of the natural vegetation that tends to colonize the abandoned walls constitute a risky situation, in particular, for the roof stability, suggesting that interventions of waterproofing of the rocky roof surroundings and the elimination of the dangerous roots that open cracks in the stone could help stabilize the whole site. The MM investigation was able to detect the points of the ceiling that showed wider vibrational displacements than the others. As such points are located close to the largest cracks in the stone an indication could be provided that these points would need attention and should be monitored carefully. The investigation through the HMI technique permitted to obtain the spectral reflectance from fragments of residual pigments on samples of selected paintings. The readability of the whole painted surface of the analyzed cycle was improved in order to identify, isolate, and study in detail all the figurative elements that comprise the scene. Such relevant contribution to the readability of the painted scene confirmed the validity of the investigation methodology in the direction of providing a valuable contribution to the documentation and to the diagnosis of the painted walls of the Tomba Campana. Despite the relevance of the contributions from the several noninvasive investigations applied to the Campana tomb, the achieved documentation is still to be considered partial and insufficient for a complete comprehension and the appropriate conservation of such an important archaeological site. In particular, long-term conservative strategies for future generation are needed. For the above reasons, site managers greatly need complete diagnostic programs and frequent monitoring of such monuments with a twofold objective: on the one side, it is essential to obtain information about the state of conservation of the tombs that can be reutilized in the future, and, on the other side, it is necessary to improve the readability also of the remaining figurative elements still to be investigated to complete the whole painted surface in order to comprehend the entire cycle in its completeness. In recent years, noninvasive techniques and 3D digital technologies, along with the structural characterization and multispectral imaging, have become common tools for the study in the field of the conservation of the cultural heritage, as they provide the most complete information to characterize the objects creating novel shared knowledge without material interaction and related risk of damage to the studied object. In these terms, the multidisciplinary tests presented on the case study of the Tomba Campana constituted a successful example that provided valuable information useful for archaeologists and conservation experts, providing a broader picture of the creative potential and technological skills applied to Etruscan culture. Acknowledgments This work was realized in the framework of the COBRA regional project “Sviluppo e diffusione di metodi, tecnologie e strumenti avanzati per la COnservazione dei Beni culturali, basati sull’applicazione di Radiazioni e di tecnologie Abilitanti” (LR 13/2008, project n. 1031); Regione Lazio is gratefully acknowledged for funding the project. A special thanks to the Soprintendenza Archeologia Belle Arti e Paesaggio per l’Area Metropolitana di Roma, la Provincia di Viterbo e l’Etruria Meridionale, and the Parco Naturale Regionale di Veio, which permitted the access to the Blue Daemons Tomb and supported the authors for the logistic issues.

1848

I. Roselli et al.

References 1. Brocato P (2008) Osservazioni sulla tomba delle Anatre e sulla più antica ideologia religiosa etrusca. Ocnus 16:69–106 2. Brocato P (2012) La tomba delle Anatre di Veio, Collana del Dipartimento di Archeologia e Storia delle Arti 5 3. Boitani F (1982) Veio: nuovi rinvenimenti nella necropoli di Monte Michele, Archeologia della Tuscia 1:95–103, tavv. XXXII:XXXIX 4. Boitani F (2010) Veio, La tomba dei Leoni Ruggenti: dati preliminari. In: Gianfrotta PA, Moretti AM (a cura di), Archeologia nella Tuscia, Daidalos 10, Atti dell’incontro di Studi (Viterbo 2007), 23–47 5. Boitani F, Neri S, Biagi F (2008) Rilessi della ceramica geometrica nella più antica pittura funeraria veiente. In: International Congress of classical archaeology meetings between cultures in the Ancient Mediterranean, Bollettino di Archeologia on line 10 (Special Volume), 20–27 6. The use of the sites is carried out by the staff of the Veii Regional Park Authority in agreement with the Soprintendenza Archeologia Belle Arti e Paesaggio per l’Area Metropolitana di Roma, la Provincia di Viterbo e l’Etruria Meridionale 7. Bartoloni G, Acconcia V, Belelli Marchesini B, Biagi F, Cerasuolo O, Neri S, Pitzalis F, Pulcinelli L, Sarracino D (2013) Progetto Veio: novità dalle ultime campagne di scavo. Scienze dell’Antichità 19(1):133–156 8. Campana GP (1843)Cenni sulla scoperta di una antica tomba etrusca presso l’antica Veio, dall’Album X:3–8 9. Campana GP (1843) Tomba di Veio. Bullettino dell’Instituto di Corrispondenza Archeologica 15:99–102 10. Canina L (1847) In: Canina (ed) L’antica città di Veii. Dai tipi dello stesso, Roma 11. Piroli L (1843) L’Album, X, 52 12. Micali G (1844) Monumenti inediti a illustrazione della storia degli antichi popoli italiani. Coi tipi della Galileiana, Firenze 13. Dennis G (1848) The cities and cemeteries of Etruria. Cambridge University Press, London 14. Cristofani M, Zevi F (1965) La Tomba Campana di Veio. Il corredo, Arch CIXVII, pp 1–35, tavv. I:XIII 15. Cristofani M (1969) In: Olschki LS (ed) Le tombe di Monte Michele nel Museo archeologico di Firenze. Olschki, Firenze 16. Roncalli F (1979) Appunti sulle ‘urne veienti’ a bauletto, Nuovi Quaderni dell’Istituto di Archeologia dell’Università di Perugia (in onore di F. Magi), 157–167 17. Verburg E (2019) The Tomba Campana: a long-debated ‘discovery’. Considering the finds of a 19th-century excavation that never happened, Kleos -Amsterdam Bull Ancient Stud Archaeol 2: 44–62 18. Banti L (1970) Le pitture della Tomba Campana a Veio. Studi Etruschi 38:27–43 19. Bettini C, Giacobini C, Marabelli M (1977) Gli ipogei dipinti della necropoli di Veio: Indagine sullo stato di conservazione e sulle tecniche pittoriche. Studi Etruschi 45:239–257 20. Harmon AM (1912) The paintings of the Grotta Campana. Am J Archaeol 16:1–10 21. Delpino F (1984–1985) Sulla scoperta della Tomba Campana di Veio: un falso dell’archeologia romantica? Rend Pont Acc 57:191–201 22. Delpino F (2012) La Tomba Campana e la sua ‘scoperta’. In: van Kampen I (ed) Il nuovo Museo dell’Agro Veientano a Palazzo Chigi a Formello. Quasar Formello 23. Athanasopuolos GA, Pelekis PC (2000) Ground vibrations from sheetpile driving in urban environment: measurements, analysis and effects on buildings and occupants. Soil Dyn Earthq Eng 19:371–387 24. Hunaidi O, Guan W, Nicks J (2000) Building vibrations and dynamic pavement loads induced by transit buses. Soil Dyn Earthq Eng 19:435–453 25. Xu YL, Hong XJ (2008) Stochastic modelling of traffic-induced building vibration. J Sound Vib 313:149–170

64

Study of Etruscan Tombs Using a Multidisciplinary Approach: Case of Campana. . . 1849

26. Singh PK, Roy MP (2010) Damage to surface structures due to blast vibration. Int J Rock Mech Min Sci 47:949–961 27. Albert DG, Taharzadeh S, Attenborough K, Boulanger P (2013) Ground vibrations produced by surface and near-surface explosions. Appl Acoust 74:1279–1209 28. Lopes P, Alves Costa P, Ferraz M, Calçada R, Silva Cardoso A (2014) Numerical modelling of vibrations induced by railway traffic in tunnels: from the source to nearby buildings. Soil Dyn Earthq Eng 61–62:269–285 29. Rainer JH (1982) Effect of vibrations on historic buildings: an overview. Assoc Preserv Technol Bull 15:2–10 30. Clemente P, Rinaldis D (1998) Protection of a monumental building against traffic-induced vibrations. Soil Dyn Earthq Eng 17:289–296 31. UNI 9916, Criteri di misurazione e valutazione degli effetti delle vibrazioni sugli edifici (in Italian), 2014, Italian standard 32. DIN 4150, Erschütterungen im Bauwesen – Einwirkungen auf baulichen Anlagen, (in German), 2015, German standard 33. SN 640312 a, Effet des ébranlements sur les constructions, (in French), 1992, Swiss standard 34. Roselli I, Malena M, Mongelli M, Cavalagli N, Gioffrè M, De Canio G, de Felice G (2018) Structural health monitoring by ambient vibration testing of the “Ponte delle Torri” of Spoleto during the 2016–2017 Central Italy seismic sequence. Int J Civil Struc Health Monitor 8(2): 199–216 35. De Silva CW (2005) Vibration and shock handbook. CRC Press, Boca Raton 36. Nakamura Y (1989) A method for dynamic characteristics estimation of subsurface using microtremor on the ground surface. Q Rep Railway Tech Res Institute 30(1): 25–33 37. Gallipoli MR, Mucciarelli M, Vona M (2009) Empirical estimate of fundamental frequencies and damping for Italian buildings. Earthq Eng Struct Dyn 38(8):973–988 38. Wadhwa N, Wu H, Davis A, Rubinstein M, Shih E, Mysore GJ, Chen JG, Buyukozturk O, Guttag JV, Freeman WT, Durand F (2017) Eulerian video magnification and analysis. Commun ACM 60(1):87–95 39. Yu-Wu H, et al. Eulerian video magnification for revealing subtle changes in the world. https:// people.csail.mit.edu/mrub/papers/vidmag.pdf. Accessed on 2 Mar 2017 40. Fioriti V, Roselli I, Tatì A, Romano R, De Canio G (2018) Motion magnification analysis for structural monitoring of ancient constructions. Measurement 129:375–380 41. De Canio G, Mongelli M, Roselli I (2013) 3D motion capture application to seismic tests at ENEA Casaccia research center: 3DVision system and DySCo virtual lab. WIT Trans Built Env 134:803–814 42. Fioriti V, Roselli I, Tatì A, De Canio G (2017) Motion magnification for urban buildings. Proceedings of 12th international conference on critical information infrastructures security, Lucca, Italy 43. Colantonio C, Pelosi C, D’Alessandro L, Sottile S, Calabrò G, Melis M (2018) Hypercolorimetric multispectral imaging system for cultural heritage diagnostics: an innovative study for copper painting examination. Eur Phys J Plus 133:526 44. Melis M, Miccoli M, Quarta D (2013) Multispectral hypercolorimetry and automatic guided pigment identification: some masterpieces case studies. In: Pezzati L, Targowski P (eds) Proceedings of SPIE 8790, Optics for Arts, Architecture, and Archaeology IV, Munich, vol 8790. SPIE, Washington, DC

The Restoration of the Acropolis of Athens: A Holistic Approach

65

Maria Ioannidou, Dorina Moullou, and Dimitris Egglezos

Contents 65.1

65.2

65.3 65.4

65.5

65.6

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.1.1 The Acropolis Hill and the Monuments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.1.2 The Structural System of the Acropolis Monuments and Their Seismic Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.1.3 The Damages Suffered by the Monuments and the Reasons for the Intervention . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Current Restoration Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.1 Organization of the Works . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.2 The Principles of the Interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.3 Implementation of Restoration Interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.4 Construction Site Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.5 Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.6 Decision-Making . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.2.7 Dissemination of Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Research and Innovation in the Implementation of the Restoration Project . . . . . Development of Innovative Methods and Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.4.1 Crack Repair in Columns of the Parthenon Opisthonaos In Situ . . . . . . . 65.4.2 Cleaning of the Surfaces with a Customized Laser System . . . . . . . . . . . . Geometric Documentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.5.1 Photogrammetry and 3D Laser Scanning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.5.2 Geophysical Prospection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diagnostic Techniques for the Surface of the Monuments . . . . . . . . . . . . . . . . . . . . . . . 65.6.1 Research on the Soot Deposits and Black Encrustations on the Marble Surface of the Monuments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.6.2 Investigation of the Polychromy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1852 1852 1858 1862 1866 1866 1868 1870 1875 1877 1878 1878 1879 1880 1880 1880 1881 1881 1883 1884 1884 1884

M. Ioannidou Director Emerita Acropolis Restoration Service, Athens, Greece D. Moullou (*) Hellenic Ministry of Culture and Sports, Athens, Greece e-mail: [email protected] D. Egglezos Independent Researcher, Athens, Greece © Springer Nature Switzerland AG 2022 S. D’Amico, V. Venuti (eds.), Handbook of Cultural Heritage Analysis, https://doi.org/10.1007/978-3-030-60016-7_65

1851

1852

M. Ioannidou et al.

65.7

1884 1886 1886 1887

Instrumental Health Monitoring and Seismic Recording . . . . . . . . . . . . . . . . . . . . . . . . . 65.7.1 Installation of an Accelerographic Array . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.7.2 Installation of Seismographs on the Parthenon . . . . . . . . . . . . . . . . . . . . . . . . . 65.7.3 Installation of a System of Optical Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.7.4 Highly Accurate Measurements of Topographical Targets on the Circuit Wall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.8 Research on the Stability and Rehabilitation of the Monuments . . . . . . . . . . . . . . . . 65.8.1 Development of a Novel Calculation Method for the Restoration Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.8.2 Research on Proper Numerical Modeling of Dry Masonry Structures . . . 65.8.3 Structural Efficiency Assessment of Restored Areas . . . . . . . . . . . . . . . . . . . 65.8.4 Assessment of the Thermal Impact on the Mechanical Properties of the Marble . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.8.5 Assessment of the Structural Pathology of the Monuments via Staged Historical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.8.6 Design and Evaluation of Mortars for the Joining Marble or Poros Architectural Blocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65.9 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1887 1888 1888 1889 1890 1891 1891 1893 1893 1894

Abstract

The Acropolis monuments (the Parthenon, the Erechtheion, the Propylaia, the Temple of Athena Nike, and the Circuit Wall) survived until today through centuries of changes in use and in form. During their long history, their continuous exposure in the action of damaging factors, both natural and human, has provoked major or minor failures related to mechanical, chemical, and biological causes. Those damages have to be confronted and thus a restoration project begun in 1975 and is continued until today. This paper presents the aforementioned restoration project, the principles, and methodology applied as well as the organization of the works (decision making, implementation, documentation, dissemination) and the disciplines involved. It focuses on the contribution of research and modern technologies in its implementation: the systematic documentation, the instrumental monitoring and recording (structural behavior, deformations, and minor movements), and the research on the stability of the monuments and on the development of innovative applications during the course of the works.

65.1

Introduction

65.1.1 The Acropolis Hill and the Monuments The Acropolis, a steep hill that rises (altitude 157 m) from the plain of Attica, was from the very beginning a natural landmark of Athenian topography (Fig. 65.1). The hill is sheer on all sides except for the west and has an almost flat top that covers an area of about 30,000 m2. Its naturally fortified position and the existence of sources of drinking water on its slopes attracted settlers as early as the Neolithic period (5500–3200 BC) and continued to be used in the centuries that followed. In the

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1853

Fig. 65.1 The Acropolis hill from the west

thirteenth century BC, the summit was protected by strong fortification walls encompassing the residence of the local Mycenaean ruler. In the following years, the hill gradually acquired a religious character and the cult of Athena, the city’s patron goddess was established. The most important time in the history of the Sacred Rock is, unquestionably, the fifth century BC. Immediately after their victory over the Persians, Athenians built a new Circuit Wall on the hill using many architectural members from the recently destroyed buildings. Shortly after, under the leadership of Perikles, a visionary politician, they carried out an ambitious building program whose purpose was to increase the splendor of Athens and, through this, to establish its political power. Four buildings of incomparable beauty were erected on the Acropolis rock: The Parthenon, a temple of the Doric order, of particularly large proportions (69.5  30.9 m measured on the stylobate), dedicated by the Athenians to the martial Athena (447–432 BC, attributed to the architects Iktinos and Kallikrates and to the sculptor Pheidias (Fig. 65.2). It is considered as an outstanding achievement of ancient Greek architecture and sculpture. It is a peripteral temple with an amphiprostyle cella divided into two interior spaces and with many Ionic features, the most important of which is the Ionic frieze. The cella used to hold the famous chryselephantine statue of the goddess. The Propylaia, the monumental entrance building to the sanctuary, famous already in antiquity for its impressive coffered ceilings (437–432 B.C.), was the work of

1854

M. Ioannidou et al.

Fig. 65.2 The Parthenon in 2011

the architect Mnesikles (Fig. 65.3). It consists of a central building with two six-columned Doric porticos east and west, and two Ionic halls in the interior, framed at the north and south by two wings with porches in the Doric order (Entire width: 48 m; central structure: 18 m  25 m). Due to the outbreak of the Peloponnesian War (431–404 BC), the building was never completed. The Erechtheion, the monument depository of the most ancient Athenian traditions, built between 421 and 406 BC, is an Ionic prostyle building (Fig. 65.4). Its unusual plan, unique for a Greek temple, was imposed by the attempt to house many different cultic areas (22.22  11.16 m measures the rectangular section of the cella). In the west part of its south side stands the famous porch of the Caryatids (Porch of the Maidens), so-named from the korai supports that hold up the roof. The entrance to the west section from the north comprises a propylon with a brilliant coffered ceiling and magnificent doorway. The Temple of Athena Nike, a small temple (8.27 m  5.64 m) built between the years 427 and 424 B.C., southwest of the Propylaia, on the bastion that fortified the southwest projection of the Sacred Rock, was designed by the architect Kallikrates (Fig. 65.5). Ionic in style, amphiprostyle, with rich sculptural decoration, it replaced an earlier, poros shrine of the goddess, which was discovered in 1936 beneath the marble temple and is preserved with the Mycenaean bastion on which it was founded, in a subterranean crypt that was created expressly for that purpose inside the classical tower.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

Fig. 65.3 The Propylaia east stoa in 2010

Fig. 65.4 The Erechtheion from southwest

1855

1856

M. Ioannidou et al.

Fig. 65.5 The temple of Athena Nike from northeast

The aforementioned buildings, built entirely of white marble, are characterized by ingenious planning, richness, and perfection of their forms as well as flawless construction. Other cult buildings, many dedicatory works of art and numerous inscriptions fill the picture of a living and active sanctuary, which had a strong appeal to the religious sentiment of the Athenians. The Acropolis remained as a functioning sanctuary down to the fourth century AC. From the end of the third century AC, however, the west side of the hill is fortified. With the domination of the new religion – Christianity – significant changes are made to the Acropolis and to its monuments. The Parthenon, whose fame has the greatest radius, is transformed into a church, a change demanding alteration of the architectural plan (sixth century AC). The same happened for the Erechtheion, which is converted into a Christian basilica and for the southwest wing of the Propylaia, likewise transformed into a church, to accommodate the new religion [1]. An important point in the history of the place is the era of Latin domination (1205–1458). During this time the Acropolis passed successively through the hands of the Franks, the Catalans, and the Florentines. Extensive fortification works were intended to strengthen the citadel’s defensive power, while the Propylaia was transformed into a fortified residence for the local ruler. In 1458 the Acropolis changes hands and falls under Ottoman control. The new administrative and

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1857

religious requirements introduce changes to the monuments. The Parthenon becomes a mosque but continues practically undamaged until 1687, when it is bombarded by the Venetians, during a siege. The explosion of the gunpowder that was stored in the building blows it up and causes its transformation from a functioning edifice to a ruin [2]. Before Morosini’s bombardment the Ottomans dismantled the Temple of Athena Nike and used its members in a new defensive fortification that could accommodate canons in front of the Propylaia on the west side of the hill [1]. The picture of the catastrophe is to be completed at the beginning of the nineteenth century with the systematic and violent removal of the sculptures of the Parthenon by the cohorts of Lord Elgin. The Acropolis is to be found in the center of events during the Greek War for independence (1821–1830). After the liberation of Greece from Ottoman domination, the Acropolis is to devolve to the new Greek State (1833), and, having pursued a long course closely connected with the history of the land, it will be promoted as the national monument par excellence. The establishment of the Greek state marks the beginning of a new epoch for the monuments of the Sacred Rock. The Acropolis becomes a reference point for the inhabitants of the newly established state and a point of recognition of Greece by Europe. Restoration of the Acropolis monuments is promoted as a national goal of great significance that is never abandoned throughout the course of modern Greek history, even when economic conditions are extremely unfavorable [3]. After 1833, when the hill comes into the hands of the Greeks, all the military installations are removed, and the Acropolis begins to function purely as an archaeological site. After the first cleaning operations, the purpose of which was to remove the late classical and medieval remains and to display the monuments of the classical period, excavations begun on a limited scale together with restoration of the monuments, empirically but with undiminished enthusiasm. The period from the end of the nineteenth century to the Second World War was characterized by extensive excavations (1885–1890) and the implementing of an ambitious program that included the restoration of all the standing monuments on a large scale (Parthenon 1898–1902 and 1923–1930, Erechtheion 1902–1909, Propylaia 1909–1917, Temple of Athena Nike 1935–1940). This program gave the monuments the form in which we know them today. Yet the practice of incorporating iron reinforcements in the architectural members of the monuments, together with the use of scattered ancient fragments as an ordinary building material, created serious problems [4, 5]. After the Second World War, the interventions on the monuments were limited. The main characteristic of that time was the recognition of the problems created by the earlier restorations and the realization that they must be confronted immediately. An all-inclusive intervention based on clear theoretical and high scholarly criteria was proposed as urgent. The establishment of the Committee for Conservation of the Acropolis Monuments in 1975 inaugurated a new era for the Acropolis and its monuments.

1858

M. Ioannidou et al.

65.1.2 The Structural System of the Acropolis Monuments and Their Seismic Performance The standing monuments on the Acropolis, namely the Parthenon, the Propylaia, the Erechtheion, and the temple of Athena Nike, are built of marble stones in the form of rectangular blocks or drums, placed one next to or above the other, without mortar (Fig. 65.6). Extraordinary precision in cutting and carving the surfaces of the stones and in fitting them together is the basic characteristic of their structure. Although constructed of separated architectural members, their “dry masonry” joining has been done so accurately that in some cases the joins are imperceptible, giving the optical impression that the construction is continuous. The marble blocks are joined to each other with metal connecting elements as the majority of monumental buildings of the Greek classical period. The joining elements are horizontal –named clamps – connecting adjacent blocks in the same course, or vertical to the layers, named dowels, connecting blocks of successive courses. They are made of iron, placed in cuttings in the marble, and completely sheathed in molten lead. The lead ensures total mechanical cohesion between the joining element and the marble, being a softer and more yielding material, and absorbs part of the shock and energy of an earthquake. Moreover, molten lead protects the iron of the joining elements from rusting by isolating them from the atmosphere (Fig. 65.7). The connecting elements contribute to the general resistance of the construction, especially against seismic load or deformation due to various other disturbances (strong movements, foundation settlements, etc.). The bearing system of the monuments depends on the perfect fitting of the stones and the joints between them. The joints are surfaces of discontinuity in the construction that withstand compressive and shear stresses but not tensile. These joins can

Fig. 65.6 The West side of the Propylaia (façade). (Drawing T. Tanoulas)

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1859

Fig. 65.7 The way of construction of a marble wall. Left, joining of two blocks of the same course with a clamp, right, joining of two blocks of successive course with a dowel

open with a resulting concentration of compressive and shear stress along an edge. This fact can result in a nonlinear behavior of the bearing system. The existence of joins affects even more the behavior of the monument when subjected to seismic force. The discontinuity articulations determine the distribution of stresses such as the size and distribution of seismic force. The articulated elements of the monument (columns, pilasters, walls) can oscillate or they can slide, thus either absorbing the force or concentrating these stresses at the ends of the blocks so that the edges break [6]. It is characteristic that the monuments of the Acropolis as structural systems meet modern requirements for seismic design, although the dimensioning of their structural elements was determined on the basis of morphological criteria, without any preliminary structural calculation in the modern sense of the term. As it can be seen in the Parthenon plan (Fig. 65.8), they are remarkable for the simplicity and distinctness of their structural function. They are characterized by the regularity of plan; they have an approximately symmetrical arrangement of their bearing elements and mass and an almost even distribution of stiffness. The tremendous stiffness of the walls in combination with the diaphragm function of the ceilings and roof by means of friction contributes to the resistance of the building to horizontal stress. Furthermore, the fact that the buildings are (totally or for the most part) founded on solid rock as well as the good design and quality of their foundations ensure their

1860

M. Ioannidou et al.

Fig. 65.8 Plan of the Parthenon. (Drawing by M. Korres)

excellent seismic behavior [7, 8]. It is notable that a great part of the damage we are obliged to face today is due not to the forces of nature, but to the activity of man. Much of the extensive damage to the monuments stems from interventions of the recent past that were intended to protect the monuments. In the future, the only strong mechanical strains are expected to be seismic, since we certainly hope that there will no longer be damage inflicted by mankind. It is therefore imperative to evaluate the efficiency of the monuments in seismic activity, taking into account the damage they have suffered through their long history, in order to make the necessary interventions. The seismic performance of the monuments is very complex and shows high nonlinearity. The intricacy of the problem is intensified, not only by the number of joins, cracks, deformations, displacements, and failure of the connecting elements. In recent years the exceedingly complicated phenomenon of the response of articulated structures to seismic action has been studied on the basis of modern scientific knowledge, through analytical and experimental methods [9–11]. From the quantitative point of view, calculating the composite motion of the hundreds of members of a classical monument under seismic load is a very difficult problem. The monument itself provides, however, many indications of its seismic behavior as does the information stamped upon it over time: the deformations it has undergone during the long centuries of its history are of themselves the most significant source for its seismic behavior. It is exceedingly important to evaluate the structural efficiency of the monument in its present state, taking into account the damage it has suffered. The monuments of the Acropolis, indeed, have shown excellent seismic resistance during the earthquakes that have struck Athens. Given this satisfactory seismic behavior up to now, interventions are carried out with respect to the original structural system. The basic principle of planning interventions on the architectural members is the restoration of the bearing capacity of each member so that it can withstand the greatest possible load. In case of overload, the joining elements are planned to absorb the seismic force without damage to the marble.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1861

From the monuments of the hill, only the Circuit Wall has a different structural system and this is mainly because of its different usage (a typical case of a retaining gravity wall in full operation from ancient times until today) and because of the numerous interventions and repairs during its lifetime, made in different time periods. The Circuit Wall, dated in the fifth century BC, has a polygonal perimeter of about 750 m and a height ranging from 4 to 18 m. It was founded directly on the rock but it does not follow a single plan since it was constructed in sections, in different phases and with different construction forms. The North Wall, known as the “Themistokleian Wall,” is thought to be erected by the Athenians after the Persian Wars (479 BC). Measuring 320 m long and 4–6 m high, it is built with dry masonry. The construction is interrupted in various points by the reuse of marble and poros stone architectural members of earlier buildings that stood on the Acropolis before the Persian invasion (Fig. 65.9) [12]. The South Wall known as the “Kimoneian wall” was built in 467 BC. The wall measures 295 m long and is constructed in two legs: the east (165 m) and the west (130 m), which meet at an angle of 152.5 . The height of the wall ranges from 10 to 18 m. It consists of orthogonal stone blocks with typical dimensions: 1.30 m  0.65 m  0.50 m (Fig. 65.10). Over the years strong buttresses, thick revetments, extensive later repairs with rubble masonry, and locally a new mortar facing (SE area) hide the original masonry from view [13]. The East Wall is the part of the wall that connects the South and the North wall. Its length is approx. 100 m, and its height ranges from 10 to 18 m. This part of the Wall collapsed during the earthquake of 1705 and was rebuilt [13]. Therefore, three different structural systems can be recognized:

Fig. 65.9 Orthophoto of the North Wall

1862

M. Ioannidou et al.

Fig. 65.10 The South and the East Walls. View from southeast

– Dry masonry, mainly in the North Wall (the authentic classical construction). No metal elements are joining the blocks. The stability of the wall against sliding depends totally on friction. – Masonry with mortar in the East Wall (the rebuilt area of the wall after the collapse caused by the 1705 earthquake). – Mixed structure, with the core of the wall built with dry masonry (with no metal joints) covered with a facade made of masonry with mortar, in the South wall. Besides the above forms of the Circuit Wall, some other structural peculiarities are encountered, composing a very complex structural system: the higher part of the eastern wing of the South Wall and the southern part of the East Wall (south: 130 m and east: 30 m long) consists of a massive huge platform (crepis) with a width of 6 m and a height of 3 m (Fig. 65.11).

65.1.3 The Damages Suffered by the Monuments and the Reasons for the Intervention The perfection of ancient building technique assured the strong resistance of the monuments to natural forces for a long time. They have survived for almost

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1863

Fig. 65.11 The inner side of the South Wall during the excavation of the years 1885–1890

25 centuries through wars, explosions, bombardments, fires, thefts, and interventions as well as alterations connected with different usage. As we have seen, human activity was the main reason that upset the balance of the ancient construction, leaving the monuments more susceptible to damage by the forces of nature. Most of the damage suffered by the monuments and now being confronted in the framework of the contemporary interventions falls into the following categories [14]: A. Mechanical: this includes damage caused to the marble of the monuments by earthquakes, explosions, bombardments, conflagrations (thermal fracturing), and ice (freezing), as well as by the expansion due to the swelling of rusting iron reinforcements used in the interventions of the recent past (Fig. 65.12). In the case of the Circuit Wall, the main damages concern permanent deformations due to soil pressures (especially in case of earthquakes). B. Chemical: this comprises forms of erosion suffered by the marble mainly as a result of acid rain, which contains solutions of various acids formed from the corresponding oxides in the atmosphere when the water vapors condense. A type of damage connected with atmospheric pollution concerns encrustations and deposits in the form of powder, carbon (black soot), and metal rust (Fig. 65.13). C. Biological: these are forms of corrosion caused by organic agents such as lichens, molds, bird droppings, plant roots, and their discharges (Fig. 65.14). Additionally, past restoration choices that are not in accordance with contemporary restoration principles have to be corrected. For example:

1864

Fig. 65.12 Mechanical damage. Joint and marble cracks

Fig. 65.13 Black soot on a slab of the Parthenon Frieze

M. Ioannidou et al.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1865

Fig. 65.14 Biological damage. Microbial presence under marble surface exfoliations

• Fragments of architectural members, whether or not they actually belonged together, were joined in order to create restorable members. To make such joins possible, broken surfaces were sawn to make them smooth. • Marble pieces or even whole blocks of the monuments were restored in wrong places. • Iron reinforcements (large or small) were incorporated within the architectural members of the monuments and encased them in cement, following the contemporary belief that this would counter the problems of their corrosion. Unfortunately, the corrosion and swelling of the iron caused the marble to break and pieces of it to fall. Therefore, the current restoration program, besides confronting the mechanical, chemical, and biological damages on the monuments aims at: • Restoring the fragments of members of the monuments and stabilizing those areas that show reduced structural efficiency. • Correcting the joints and the errors in placing of the architectural members. • Identifying and including in a monument’s restoration scattered members that had not been used in previous restorations, so as to save the members per se, as well as to enhance the structural stability of the monument.

1866

65.2

M. Ioannidou et al.

The Current Restoration Project

The current restoration project has been carried out since 1975 under the scholarly supervision of the “Committee for Conservation of the Acropolis Monuments” (ESMA). The ESMA is an interdisciplinary committee of experts in various fields (archaeologists, architects, civil engineers, chemical engineers, and conservators) with the task of studying the serious problems of the Acropolis monuments and intervening in order to tackle the extremely serious problems threatening them. Members of the ESMA are the heads of the services of the Hellenic Ministry of Culture and Sports (ex officio) as well as specialists, scholars, and scientists from academia. It must be stressed that all the ESMA members during the 44 years of its function have been working gratis. The implementation of the ESMA’s decisions had been carried out until 1999 by a technical office functioning within the context of the then Acropolis Ephorate of Antiquities, and since 2000 by a specifically established service of the Hellenic Ministry of Culture and Sports, the Acropolis Restoration Service (YSMA). YSMA’s purpose is the organizing and carrying out the restoration project. Indeed, the establishment of the YSMA gave the works the necessary impetus and facilitated the whole project in numerous ways. The ESMA was reestablished with a purely scholarly role and prerogatory regulations were enacted for economic matters, the salaries of the personnel, and a significant increase in the number of scientific and technical staff. Consequently, all the studies for the restoration of the monuments as well as their implementation are conducted by the YSMA highly qualified, scientific, and technical personnel (archaeologists, architects, civil, chemical and mechanical engineers, conservators, surveyors, marble cutters, specialized technicians, administrative personnel). The works are executed with the method of direct labor, which means that the service recruits the necessary personnel to both study and implement the works; very specialized studies and works supporting the project are outsourced or executed in collaboration with the YSMA personnel. For restoration works and especially for works of such extent, diversity, complexity, and significance as the ones on the Acropolis monuments, direct labor is the only suitable approach. Therefore, the structure of the YSMA has been meticulously studied in order to meet all the needs of the Acropolis restoration project.

65.2.1 Organization of the Works Structure of the YSMA The director of the YSMA, architect or civil engineer, is responsible for the coordination, scheduling, management, and supervision of the works, for its adherence to the time-frame and for introducing to the ESMA matters for which the committee is responsible, such as the scientific studies, proposals, and technical issues of the restoration works.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1867

The YSMA has been organized in the following sections: – Technical office and worksite for conservation and restoration of the Parthenon – Technical office and worksite for conservation and restoration of the Propylaia – Technical office and worksite for conservation and restoration of the Temple of Athena Nike – Technical office and worksite for conservation consolidation and restoration of the Circuit Wall of the Acropolis – Technical electromechanical office – Office for inventorying, documenting, and organizing the scattered architectural members – Office for conservation of the surface – Laboratory of casts – Photographic laboratory – Documentation office, including the database, plan, and photographic archives and library – Office of information and education – Office of the secretariat – Accounting office – Office of support and allotment of material, including the storeroom and services of a general nature

Funding of the Interventions It is worth noting that from the very beginning of the interventions until 1999, funding was assured from the Greek State together with a small contribution by the European Union. With the establishment of the YSMA and the simultaneous expansion of the interventions, the Acropolis works were included in the third Community Support Framework (2000–2006) and the National Strategic Reference Framework (2007–2017) and thus co-financed by the Greek State and the European Union. Projects: Completed and Under Completion Since 1975, a number of restoration projects have already been completed and entire monuments, or areas of monuments, have been returned to the public. The list comprises the following: • Restoration of the Erechtheion (1987) • Consolidation of part of the rocks of the slopes of the Acropolis hill (1992) • Restoration of the Parthenon: east side (1993), pronaos (2004), opisthonaos (2004), north side (2010), west side (2014) • Restoration of the Propylaia: east side, the South Wall and the superstructure of the central building (2009), south wing (2015), northwest corner (2015) • Restoration of the temple of Athena Nike (2010)

1868

M. Ioannidou et al.

A number of restoration projects are still under completion. These are: • Consolidation of parts of the circuit walls • Consolidation of the remaining unconsolidated part of the rocks of the slopes of the hill • Restoration of the Parthenon: North Wall of the cella and west side The project of the restoration of the Erechtheion was awarded the European Union Prize for Cultural Heritage/Europa Nostra Award in 1988. In 2002 marble-cutter technician Yiannis Arbilias head of the crew of the Parthenon restoration program from 1983 to 2000 was awarded the Medal of Honor of Europa Nostra. In 2013 the Acropolis works were awarded two prizes for cultural heritage by Europa Nostra: the Committee for the Conservation of the Acropolis Monuments was awarded the Grand Prix in category 3 (dedicated service) for the total of its work and “dedicated service in planning and guiding the conservation of one of the most iconic sites of European culture.” The restoration of the Propylaia, moreover, won the award in category 1 (conservation) as a work “enhancing the inherent formal and social values of an iconic monument.”

65.2.2 The Principles of the Interventions The interventions of the Acropolis monuments are pervaded by the principles of the Venice Charter (1964), which constitutes the first internationally accepted framework of principles wherein the ethics for the restoration of monuments has been codified. The interdisciplinary approach to the interventions during the phases of study and application of the works, the publication of general and specialized studies for the restoration of the monuments prior to the actual interventions, their submission as top-level studies to a procedure of multiple assessment, and then the final publication of the works follow directly the requirements of the charter. The long and arduous procedure of the development and establishment of the theoretical and scientific frame of the Acropolis restoration works, through theoretical reflections, discussions, evaluations, and confrontations, taking into account the special structural characteristics of the classical monuments, lead the ESMA to the formulation of additional principles applying to the interventions of the Acropolis monuments [15–21]. – The reversibility of the interventions, that is, the potential to reinstate the monument to the condition in which it was prior to the intervention – Respect, during the new intervention, for the authentic material and the original structural system of the monuments – Limiting interventions to the absolutely necessary ones, as far as possible – Respect and preservation of the established image of the monuments during restoration and changing of it only if the original form of the monument is certain – Total transparency throughout the works and, on completion, presentation of the works to both the scholarly world and the general public, as appropriate.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1869

These principles led gradually to specific choices in executing the works that have remained unchanged from the time of their beginning up to now. The most important of these are: • After their repair, the members are always set in the monuments in their authentic positions, which have been determined by research. The criteria for the research were measurements of the dimensions of the blocks to submillimeter precision; the quality of the marble; the geological layering; the existence of interior marble discontinuities (kommoi), color, and texture; the quality of work on the surface of the fragments, as well as deterioration features on the marble surface; and special architectural and construction details. The enrichment and deepening of our knowledge about the original architecture of the monuments, mainly our continuous and constant experience with them had brought unexpected results, such as the identification of the faulty positioning, during past restorations, of the architectural members of the monuments. This knowledge has led to the rearrangement of parts of monuments that had been restored in the past, by resetting these members in their correct places, on the basis of new information that has come to light after they were dismantled (Fig. 65.15).

Fig. 65.15 Misplacement of architectural members in the Balanos restoration. Each original column and adjacent entablature should be one color (Study L. Labrinou, Digital drawing P. Kostantopoulos)

1870

M. Ioannidou et al.

Fig. 65.16 Propylaia. Coffered slab fragments that “belong together”

• Fragments are joined only if they come from the same architectural member, that is to say, if they originally “belong together” (Fig. 65.16). • To avoid further deterioration, the original architectural sculptures are transferred to the Acropolis Museum and replaced on the monuments by exact copies of artificial cast stone, worked especially for color and texture (Fig. 65.17)

65.2.3 Implementation of Restoration Interventions The Methodology of Structural Restoration In the Acropolis monuments’ restoration, those parts of the monuments that had been restored in the past – and sometimes sections that had not been restored before but evidenced the same damage and fragmentation – are dismantled, following the articulated system of structure of classical monuments [19, 22] (Fig. 65.18). The dismantled members are restored in the laboratory. The rusted joining elements and the filling material of the previous interventions – cement mortar or rarely lead – are removed (Fig. 65.19).

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1871

Fig. 65.17 The transferring of the sixth Caryatid from the Erechtheion to the Acropolis museum

The fragments of the ancient members are rejoined. Titanium reinforcements are used for the structural restoration of the architectural members; titanium is a relatively light metal, of satisfactory strength and with properties (the coefficient of thermal expansion and the modulus of elasticity) that allow it to work well with marble. The main characteristic for which it was chosen for connecting the members, however, is its excellent resistance to all forms of corrosion. Threaded titanium rods were inserted into holes in the marble mass and secured by an inorganic plaster that was made of white cement. Fragments that do not belong together, that is to say, that do not come from the same architectural member, are never joined together (Fig. 65.20). • Where considered necessary, missing parts of the members are filled in with new Pentelic marble (same marble as the one used for the original monuments’

1872

M. Ioannidou et al.

Fig. 65.18 Dismantling the northeast corner of the Parthenon

construction) in order to restore their original form and structural independence. The new marble fillings are usually limited and the criterion for the decision is always the structural efficiency of the member and of the monument and the structural and aesthetic autonomy of the areas being restored. The aim is not to rebuild the monument but to preserve the character of the ruin and whatever bears witness to its course through history.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

Fig. 65.19 Removal of iron reinforcements from Propylaia beams

Fig. 65.20 Joining the fragments of an architrave block of the Parthenon

1873

1874

M. Ioannidou et al.

The joining surfaces of the new fillings are cut with the use of a pointing device named pondadoros or with an electric pantograph so that they agree precisely with the broken surfaces of the ancient member being filled in. The outer surfaces of the fillings and of the members made of new marble are finished by hand so as to produce precisely the ancient form and decorative aspect of each member. The goal is to restore the architectural members according to their original dimensions and to assure the function of the structural system in agreement with the building principles applied in antiquity. After being restored, the members are reset in the positions they had in antiquity, only in exceptional cases in similar positions (Fig. 65.21). They are joined with titanium sheets that are made into clamps and dowels and are anchored in the ancient sockets and cuttings with inorganic mortar corresponding to the ancient system of joining (Fig. 65.22). In this process a perfect fit is sought, an inviolable canon, that ensures the high quality of the contemporary intervention. During the resetting of members, geometrical deformations of the area that was dismantled are partially corrected to the extent allowed by the remaining distortions of the members that were not dismantled, in order to achieve as much as possible the original form. In addition to dismantled architectural members, scattered ancient members that have been identified and recognized as belonging to monuments are also reset in their original or a corresponding position. In a few cases, for reasons of stability, it

Fig. 65.21 Resetting of an architrave block of the Parthenon north colonnade

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1875

Fig. 65.22 Resetting of the titanium clamps in the ancient cuttings

may be preferable to set a few architectural members made entirely of new marble rather than using ancient members restored with new marble that is preserved in fragmentary condition (Fig. 65.23). Together with the works of structural restoration, conservation is carried out on the marble surfaces of the monuments that show damage from the action of atmospheric pollution, from biological and physiochemical agents and from human activity. These works comprise the consolidation of disintegrating surfaces, the joining of fragments, the filling of cracks and spaces with injected mortar, the removal of soot deposits and black crust, and the application of an artificial patina to new marble fillings. The methods and materials employed are based on international bibliography, the experience of the worksite, and the original research carried out in the past on Pentelic marble, and they are reversible, according to their behavior over time.

65.2.4 Construction Site Organization When restoring a monument, in addition to specifying the damage and determining of ways of confronting it, it is of vital importance to resolve the issues of choosing the ways and the technical means by which the restoration works will be carried out. Organizing a worksite for a technical work is complex in itself, but it is all the more so when it is a question of restoring a monument, since in this case the worksite comprises all the infrastructure and the materials to serve both restoration and

1876

M. Ioannidou et al.

Fig. 65.23 The central building of the Propylaia after the completion of the restoration

conservation. In most cases, the area and space available are limited, and the access and transportation of both technical equipment and supplies are difficult. The constant presence of visitors and the respect for the historical and artistic value of the monument demand the careful study of the technical infrastructure and special arrangements of the worksite [23, 24]. The basic technical means in use today for restoration projects – hoisting systems, systems of moving and transporting – are the counterpart of the ones that were used in the construction of the monuments, except the motive power used today for the systems and the choice of materials. Nevertheless, the great progress seen in the works during the past few years has been accompanied by the development of fully up-to-date technology in the interventions, in full combination with the ancient methods and techniques, used for the initial construction. The first works on the Acropolis began with tremendous difficulties: mechanical means were few, and the difficulty of transporting vehicles up to the rock and moving them on the rock meant a reliance on antiquated systems of transport based on human hands. As the works progressed, the picture of the worksites gradually changed. Innovative technical methods were applied while original technological applications and inventions accelerated the works (Figs. 65.10 and 65.24). Notable among these are: • The problem of transporting the necessary materials and massive blocks of marble to the top of the Acropolis was resolved by designing and constructing a special hoisting machine, which, when not in use, can be folded so as to be invisible from the city.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1877

Fig. 65.24 The Acropolis worksite. View from west

• The transportation of materials and large pieces of marble on the Acropolis was resolved by using systems capable of moving over the rough surface of the rock. • The hoisting machinery used for dismantling and re-constructing the monuments makes use of the developments of modern technology. Thus we have systems of cranes and bridge cranes on scaffolding that are powered by electricity at very low speeds and can move with great precision. • Innovative constructions and arrangements were designed by the YSMA engineers and were constructed for the purpose of accelerating the works. Characteristic examples are grabs for lifting architectural members, special pointing devices. Especially important among all these devices is the original machine that was devised for cutting the flutes of the column drums of new marble, accurate to the millimeter. Its use has greatly accelerated the work of restoration. The final phase of cutting the flutes is to be done by hand by the experienced marble technicians of the service.

65.2.5 Documentation During the restoration works, particular weight is attached to the documentation of the monuments’ state of preservation prior to the interventions, as well as to documenting the works during all their phases. The material produced by the

1878

M. Ioannidou et al.

systematic documentation (drawings, photographs, films, and instrumental monitoring outputs) is entered since the beginning of the restoration works into a constantly upgraded electronic database so that it can easily be accessed by researchers. Content of the database is available to the public via the YSMA repository [25, 26].

65.2.6 Decision-Making The practice of preparing and executing the studies involves systematic documentation, study, interdisciplinary approach, international dialogue among specialists, and continual development of the relevant know-how. The study is first discussed in the framework of ESMA. After the study has been approved by the committee submitted for approval to the Central Archaeological Council of the Hellenic Ministry of Culture and Sports, which gives its opinion. Thus, the project is implemented consequently upon a ministerial decision. Every study is published and sent to university and research libraries. However, prior to the implementation of each major study, in order to ensure an objective approach to decision making at a scholarly level, the ESMA has established a procedure that includes the organizing of international meetings of invited experts from all over the world, where the studies drafted for a restoration project – prior to the interventions – are presented and discussed. In this way the problem-solving process has been extended, beyond national borders. Six such meetings are held so far (1987, 1983, 1989, 1994, 2002, 2013) with great success and impact at an international level.

65.2.7 Dissemination of Information The ESMA has established a means of communicating with various interest groups (scientific community, students and educators, wider public) through the regular publication of the restoration results, at multiple levels. The organization of international meetings, as mentioned above, where the restoration studies are submitted for criticism before their application on the monument is a sine qua non of the typical procedure. Moreover, through 1-day seminars, lectures, and presentations, the most recent results are presented to the community. Finally, a number of publications, for both scholar and the wider public, together with other means of dissemination of the information (website, films, and exhibitions) keep the international community updated about the achievements of the ESMA. Particular concern has been shown for the education of the youngsters not only about the restoration works but also about the art and architecture of the buildings, as well as various aspects of the classical past. A great number of educational programs have been designed, in order to give students and educators the opportunity to approach the richness of the monuments, their qualities, and the process of their restoration.

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1879

In this framework the YSMA created a “virtual theater” of 50 seats in the new Acropolis Museum. In the theater museum guests can watch stereoscopic and simple projections of documentaries produced by the service. Besides the traditional type of documentaries, three 3D graphic documentaries have been produced (until today): entitled “The Acropolis in Antiquity,” “Building an Ancient Temple,” and “The Anastelosis of the Erechtheion.”

65.3

Research and Innovation in the Implementation of the Restoration Project

The main characteristic of the restoration project carried out on the Acropolis is the continuous research that accompanies it from the very beginning through all phases of the works. The special studies made prior to each intervention and the scholarly research by the specialist personnel in collaboration with educational or other research institutions that accompany all phases benefit the works and establish them as a reference point for all comparable projects. State-of-the-art technology is applied to the initial studies for the interventions, the organization of the worksites, the performing of the works, and to their systematic documentation. Traditional techniques and materials are utilized together with contemporary methods and materials that are compatible with those of the original monuments. The use of advanced technology in study research and the execution of the works has been a fundamental characteristic of the interventions on the Acropolis monuments, from their very beginning in 1975. Especially after 2000 and the establishment of the YSMA, state-of-the-art technology supported the expansion and expedition of the restoration program that led to innovative methods and to technological applications. Thanks to research during all these years, much new information has come to light about the archaeology, history, and architecture of the Acropolis monuments, while in these fields the monuments are – and will continue to be – the focus of study by leading scholars. The new information has not simply increased our knowledge and added to the relevant bibliography; it has contributed to the conduction of studies that are fundamental to the interventions being carried out. Yet, theoretical, scholarly, and technical problems have been encountered during the application of these studies. To resolve those problems, research has been carried out by the scholarly personnel of the works, either alone or in collaboration with educational or research institutions (e.g., National Technical University of Athens, the Aristotle University of Thessaloniki, the University of Crete, the University of Patras, the Institute of Technology and Research of Crete, the National Observatory of Athens, the Nuclear Research Center “Democritos,” the British Museum, l’ Université X de Paris, the Mie University of Tokyo, the ETH Zurich, the National Research Council of Canada, and others). Within the framework of the academic and scholarly research undertaken in recent years, the most important could be resumed as follows.

1880

65.4

M. Ioannidou et al.

Development of Innovative Methods and Techniques

65.4.1 Crack Repair in Columns of the Parthenon Opisthonaos In Situ Severe cracks in columns of the Parthenon Opisthonaos were caused by a thermal fracture in antiquity. During the restoration process, and in order to avoid dismantling the columns and to retain their authentic structure, the cracks in the drums were injected with a hydraulic grout of high penetration especially studied for this purpose. The project was particularly difficult, because the injection should be applied under low pressure and follow an upward path. For this reason, two distinct compositions of hydraulic grouts were designed: a moderate to high strength grout for the welding of the column fragments (consistent with the marble mechanical properties) and a low strength grout for the sealing of the microfractures in the vicinity of the drum joints (Fig. 65.25).

65.4.2 Cleaning of the Surfaces with a Customized Laser System The first application of this system was the cleaning from the soot deposit and the black crust of the surfaces of the dismantled blocks of the Parthenon west frieze. It was a most sensitive and delicate operation, since the historical layers of the surface of the sculpture had to be preserved. The cleaning method chosen was the laser in an entirely original application, using simultaneously two wavelengths, infrared, and ultraviolet, developed

Fig. 65.25 Parthenon. In situ repair of the cracks of the Opisthonaos columns

65

The Restoration of the Acropolis of Athens: A Holistic Approach

1881

Fig. 65.26 The laser cleaning system

by the Institute of Electronical Structure and Laser of the Technological and Research Foundation of Crete (Fig. 65.26). By means of this system, the frieze blocks were safely cleaned without affecting the underlying layer. In addition, the historical layers of the surfaces of the frieze were preserved, with an excellent aesthetic result; information about the original workmanship and traces of ancient color were also revealed.

65.5

Geometric Documentation

65.5.1 Photogrammetry and 3D Laser Scanning Geometric documentation is the first step in the restoration of a monument. What is of particular interest is the recording and mapping of the changes in the original geometry of the monument and the damage it has suffered. These features, in connection with

1882

M. Ioannidou et al.

the study of the historical and archaeological evidence and the structural pathology of the monument, form the necessary infrastructure for its restoration. In the case of the Acropolis, from 1975 on, the possibilities offered by the use of conventional methods and instruments were employed for surveying the monuments. In recent years a special emphasis has been placed on using techniques of advanced technology such as photogrammetry and 3D laser scanning for the geometric documentation of the monuments and the fortification walls and the rock itself. In 2007–2009 a project entitled “Development of a Geographical Information System in the Acropolis of Athens” was carried out [27, 28]. Its aims were to solve a number of problems connected mainly with any restoration study of the circuit Walls such as: • The full and analytical geometric documentation of the monument and the underlying rock that was inaccessible any other way • The creation of an infrastructure for: – The architectural and archaeological documentation (of the building – historical phases and for recording the incorporated members) – The determination of the pathology, the static and seismic structural behavior of the walls The deliverables or the project were: • Unified geodetic infrastructure (including old and new networks) • Orthophotomosaic of the plan view of the hill at a scale of 1:100 (pixel size 10 mm, accuracy