Encyclopedia of Sensors and Biosensors, Four Volume Set [1 ed.] 0128225483, 9780128225486

Encyclopedia of Sensors and Biosensors, Four Volume Set provides unique foundational level information on the rapidly gr

129 54 134MB

English Pages 2958 [2980] Year 2022

Report DMCA / Copyright

DOWNLOAD PDF FILE

Table of contents :
Volume 1
Volume 2
Volume 3
Volume 4
Recommend Papers

Encyclopedia of Sensors and Biosensors, Four Volume Set [1 ed.]
 0128225483, 9780128225486

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

ENCYCLOPEDIA OF SENSORS AND BIOSENSORS

This page intentionally left blank

ENCYCLOPEDIA OF SENSORS AND BIOSENSORS EDITOR IN CHIEF

Roger Narayan Joint Department of Biomedical Engineering University of North Carolina and North Carolina State University Raleigh United States SECTION EDITOR

Kazunori Ikebukuro Department of Biotechnology and Life Science, Graduate School of Engineering Tokyo University of Agriculture and Technology Koganei, Tokyo Japan and Research Center for Functional Materials National Institute for Materials Science (NIMS) Tsukuba Japan

VOLUME 1

Amsterdam • Boston • Heidelberg • London • New York • Oxford Paris • San Diego • San Francisco • Singapore • Sydney • Tokyo

Elsevier Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom 50 Hampshire Street, 5th Floor, Cambridge MA 02139, United States Copyright Ó 2023 Elsevier Ltd. unless otherwise stated. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers may always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN 978-0-12-822548-6

For information on all publications visit our website at http://store.elsevier.com

Publisher: Oliver Walter Acquisitions Editor: Clodagh Holland-Borosh Content Project Manager: Katarzyna Miklaszewska Associate Content Project Manager: Ramalakshmi Boobalan Designer: Matthew Limbert

EDITOR IN CHIEF

Dr. Roger Narayan is a Distinguished Professor in the Joint Department of Biomedical Engineering at the University of North Carolina and North Carolina State University, United States. He is an author of over 200 publications as well as several book chapters on processing of microstructure and nanostructured materials for biosensing and other medical applications. He currently serves as an editorial board member for several academic journals, including as associate editor of Applied Physics Reviews (AIP Publishing). Dr. Narayan has edited several books, including the textbook Biomedical Materials, currently in its second edition (Springer), the handbook Materials for Medical Devices (ASM International), and the Encyclopedia of Biomedical Engineering (Elsevier). He has previously served as the director of the ASM International Emerging Technologies Awareness Committee, the TMS Functional Materials Division, and the American Ceramic Society Bioceramics Division. As the 2016–17 ASME Swanson Fellow, Dr. Narayan worked with America Makes, the US national additive manufacturing institute, on activities to disseminate additive manufacturing technologies, including the development of a national workforce/education/outreach roadmap for additive manufacturing and the development of a repository containing educational materials related to additive manufacturing. Dr. Narayan has received several honors for his research activities, including the University of North Carolina Jefferson-Pilot Fellowship in Academic Medicine, the NCSU Alcoa Foundation Engineering Research Achievement Award, the National Science Faculty Early Career Development Award, the Office of Naval Research Young Investigator Award, and the American Ceramic Society Richard M. Fulrath Award. He has been elected as Fellow of the Materials Research Society, the American Ceramic Society, AAAS, ASME, and ASM International.

v

This page intentionally left blank

SECTION EDITORS Beatriz Fresco Cala is a postdoctoral researcher at Analytical Chemistry Department (University of Córdoba, Spain). Her research interests focus on the emulsions-based synthesis procedures, development of selective polymeric materials, miniaturized systems, micro- and nanofabrication, biomimetic particles, and applications in environmental and biomedical/clinical analysis. Dr. Fresco Cala studied Chemistry at the University of Córdoba, finishing her bachelor’s degree with the Extraordinary End-of-Studies Award in 2012. In 2014, she received a competitive predoctoral grant (FPU) for the development of her Ph.D. During this period, she performed short-term research stays at the University of Valencia (Spain, 2016) and at the Academy of Sciences of the Czech Republic (2017). She received her Ph.D. in Chemistry in October 2018 with International Mention and Sobresaliente Cum Laude. She received awards for her Ph.D. and for her scientific career from the Andalusian Regional Group of the Spanish Society of Analytical Chemistry (GRASEQA), the Spanish Society of Analytical Chemistry (SEQA), Lilly Company, and Royal Spanish Society of Chemistry (RSEQ). During the postdoctoral stage, she has also been a beneficiary of several competitive Spanish contracts and the prestigious international ‘Alexander von Humboldt’ fellowship for 2,5-year postdoctoral research stay in Germany. To date, Dr. Beatriz Fresco Cala is author/co-author of over 23 peer-reviewed publications and 2 book chapters, which have been cited 400 times (h-index ¼ 12). Dr. Kazunori Ikebukuro received his B.Sc. degree in 1989 from University of Tokyo in Japan, M.Sc. degree in 1992 from University of Tokyo and Ph.D. degree in 1996 from University of Tokyo. During his M.Sc. course, he received the British Council scholarship and studied at Cranfield Institute of Technology in the United Kingdom from 1990 to 1992. His Ph.D. thesis is on the development of biosensor for environmental monitoring. Dr. Ikebukuro worked at the Research Center for Advanced Science and Technology, University of Tokyo, Japan as a research associate from 1993 to 1996, and as a lecturer from 1996 to 2001. He then worked at the Department of Biotechnology and Life Science, Tokyo University of Agriculture and Technology as an associate professor and has been working as a professor at TUAT since 2009. His research interests are in 1) Development of functional aptamers for theranostics using evolutionary molecular engineering strategy, 2) Development of novel sensing technologies using combination of aptamers, enzymes and antibodies, and 3) Development of detection system for gene and its epigenetic modification focusing on DNA/RNA structures. He is always focusing on the design and development of molecular recognition device for biosensors.

vii

viii

Section Editors

Roozbeh Jafari is the Tim and Amy Leach Professor of Biomedical Engineering, Computer Science and Engineering and Electrical and Computer Engineering at Texas A&M University, United States. He received his Ph.D. in Computer Science from UCLA and completed a postdoctoral fellowship at UCBerkeley. His research interest lies in the area of wearable computer design and signal processing. He has raised more than $86M for research with $23M directed towards his lab. His research has been funded by the NSF, NIH, DoD (TATRC), DTRA, DIU, AFRL, AFOSR, DARPA, SRC, and industry (Texas Instruments, Tektronix, Samsung, and Telecom Italia). Dr. Jafari has published over 200 papers in peer-reviewed journals and conferences. He has served as the general chair and technical program committee chair for several flagship conferences in the areas of wearable computers. He is the recipient of the NSF CAREER award (2012), IEEE Real-Time & Embedded Technology & Applications Symposium best paper award (2011), Andrew P. Sage best transactions paper award (2014), ACM Transactions on Embedded Computing Systems best paper award (2019), and the outstanding engineering contribution award from the College of Engineering at Texas A&M (2019). He has been named Texas A&M Presidential Fellow (2019). He serves on the editorial board for the IEEE Transactions on Biomedical Circuits and Systems, IEEE Sensors Journal, IEEE Internet of Things Journal, IEEE Journal of Biomedical and Health Informatics, IEEE Open Journal of Engineering in Medicine and Biology, and ACM Transactions on Computing for Healthcare. He is currently the chair of the IEEE Wearable Biomedical Sensors and Systems Technical Committee (elected), as well the IEEE Applied Signal Processing Technical Committee (elected). He frequently serves on scientific panels for funding agencies, served as a standing member of the NIH Biomedical Computing and Health Informatics (BCHI) study section (2017-2021), and currently is the chair of the NIH Clinical Informatics and Digital Health (CIDH) study section (2020-2022). He is a Fellow of the American Institute for Medical and Biological Engineering (AIMBE).

Dr. Boris Mizaikoff is a Chaired Professor and Director of the Institute of Analytical and Bioanalytical Chemistry at Ulm University, Ulm, Germany. Since 2021, he is also a Director at the Hahn-Schickard Institute for Microanalysis Systems in Ulm. His research interests focus on optical sensors, biosensors, and biomimetic sensors in the mid-infrared spectral range, system miniaturization and integration based on micro- and nanofabrication, multifunctional (nano)analytical platforms, development of biomolecular/biomimetic recognition architectures, multivariate data evaluation, and applications in environmental analytics, process analysis, and biomedical/clinical diagnostics. He is author/co-author of over 400 peer-reviewed publications, 18 patents, and over 100 plenary, keynote, and invited contributions at scientific conferences.

Thiago Regis Longo Cesar da Paixão received his B.Sc. from the Institute of Chemistry of the University of São Paulo in 2001 and became a graduate student at the same institution, where he received his M.Sc. (2004) and Ph.D. (2007) under the supervision of Dr. Mauro Bertotti. For a year (2008/2009), he was a postdoctoral fellow in the group of Professor Mauro Bertotti at the same University. Following his postdoctoral fellowship, Dr. Paixão was appointed as an Assistant Professor at the University Federal of ABC where he stayed for 2 years. In 2011, he was hired as an assistant professor at the University of São Paulo and promoted to Associate Professor in 2016. At the beginning of 2018, he was nominated as an affiliate member of the Brazilian Academy of Science as a promising young researcher. Dr. Paixão’s fields of interest include chemical sensors, electronic tongues, and paper-based devices aiming at in-field applications.

Section Editors

ix

Mehmet R. Yuce received his M.S. degree in Electrical and Computer Engineering from the University of Florida, Gainesville, United States in 2001, and Ph.D. degree in Electrical and Computer Engineering from North Carolina State University (NCSU), Raleigh, United States in December 2004. He was a post-doctoral researcher in the Electrical Engineering Department at the University of California at Santa Cruz in 2005. He was a Senior Lecturer in the School of Electrical Engineering and Computer Science, University of Newcastle, NSW, Australia until July 2011, when he joined the Department of Electrical and Computer Systems Engineering, Monash University, Australia. His research interests include low-power electronics design, IoT sensor, implantable and wearable medical devices, cuffless blood pressure device design, telemedicine, wireless body area network (WBAN), biosensors, MEMs sensors and actuators, integrated circuit technology, radio frequency circuit design, and energy harvesting. Dr. Yuce has published more than 250 technical articles in these areas and received a NASA group achievement award in 2007 for developing an SOI radio transceiver. He is an author of three books in the area of wearable medical devices. He is a topical editor for IEEE Sensors Journal, an editor-in-chief for Sensors journal, and guest editor for several IEEE journals.

This page intentionally left blank

CONTRIBUTORS FOR ALL VOLUMES Hasan T Abbas James Watt School of Engineering, University of Glasgow, Glasgow, United Kingdom

Renata BA Albuquerque Fluminense Federal University, Niteroi, Rio de Janeiro, Brazil

Qammer H Abbasi James Watt School of Engineering, University of Glasgow, Glasgow, United Kingdom

Brigida Alfano ENEA CR-Portici, Portici, Italy

R Abdel-Karim Department of Metallurgy, Faculty of Engineering, Cairo University, Giza, Egypt Barbara Adinolfi Istituto di Fisica Applicata “N. Carrara”dIFAC del Consiglio Nazionale delle Ricerche, Florence, Italy Wafa Aidli Dipartimento di Chimica, Gruppo di Chimica Elettroanalitica (ELAN), Università degli Studi di Milano, Milan, Italy Semra Akgönüllü Department of Chemistry, Hacettepe University, Ankara, Turkey Deji Akinwande Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, United States B Erdem Alaca Department of Mechanical Engineering, Koç University, Istanbul, Turkey; and Surface Science and Technology Center (KUYTAM), Koç University, Istanbul, Turkey Tuncay Alan Department of Mechanical and Aerospace Engineering, Monash University, Clayton, VIC, Australia Georgina Alarcón-Angeles Universidad Autónoma Metropolitana-Xochimilco, Departamento de Sistemas Biológicos, México, Mexico Sebastián Alberti UiT The Arctic University of Norway, Tromsø, Norway

Akram Alomainy School of Electronic Engineering and Computer Science, Queen Mary University of London, London, United Kingdom Fadi Alsaleem Durham School of Architectural Engineering and Construction, Mechanical Engineering Department, University of Nebraska-Lincoln, Lincoln, NE, United States Giaan Arturo Álvarez-Romero Universidad Autónoma del Estado de Hidalgo, Ciudad del Conocimiento, Carretera Pachuca-Tulancingo, México, Mexico Aziz Amine Equipe Analyses Chimiques et Biocapteurs, Laboratoire Génie des Procédés et Environnement, Faculté des Sciences et Techniques, Université Hassan II de Casablanca, Mohammedia, Morocco Gustavo FS Andrade Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Rafaela S Andre Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, Sao Carlos, Brazil Lúcio Angnes Department of Fundamental Chemistry, Institute of Chemistry, University of Sao Paulo, Sao Paulo, Brazil

xi

xii

Contributors For All Volumes

Takahiro Arakawa Department of Biomedical Devices and Instrumentation, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo, Japan Koiti Araki Department of Fundamental Chemistry, Institute of Chemistry, University of Sao Paulo, Sao Paulo, Brazil Juan Pablo Arango GIBEC Research Group, Life Sciences Faculty, Universidad EIA, Envigado, Colombia Ryutaro Asano Department of Biotechnology and Life Science, Graduate School of Engineering, Tokyo University of Agriculture and Technology, Tokyo, Japan Larissa Bach-Toledo Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil Raphael P Bacil Department of Fundamental Chemistry, Institute of Chemistry (IQUSP), University of São Paulo, São Paulo, Brazil Senem Kurşun Bahadır Faculty of Mechanical Engineering, Department of Mechanical Engineering, Istanbul Technical University, Istanbul, Turkey Baijnath Shriram Institute for Industrial Research, Delhi, India Camelia Bala Laboratory for Quality Control and Process Monitoring, University of Bucharest, Bucharest, Romania; and Department of Analytical Chemistry, University of Bucharest, Bucharest, Romania Jenitha Antony Balasingam Department of Electrical and Computer Engineering, Faculty of Engineering, University of Windsor, Windsor, ON, Canada Francesco Baldini Istituto di Fisica Applicata “N. Carrara”dIFAC del Consiglio Nazionale delle Ricerche, Florence, Italy Ivneet Banga Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States Ganesh Chandra Banik Department of Soil Science and Agricultural Chemistry, Uttar Banga Krishi Viswavidyalaya, Cooch Behar, India

Craig E Banks Manchester Metropolitan University, Manchester, United Kingdom Ahmed Barhoum National Centre for Sensor Research, School of Chemical Sciences, Dublin City University, Glasnevin, Ireland Diandra Nunes Barreto Institute of Chemistry, Federal University of Uberlandia, Uberlândia, Brazil Kaory Barrientos GIBEC Research Group, Life Sciences Faculty, Universidad EIA, Envigado, Colombia Alex D Batista Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany Alex Domingues Batista Institute of Chemistry, Federal University of Uberlandia, Uberlândia, Brazil Krzysztof B Bec Institute of Analytical Chemistry and Radiochemistry, University of Innsbruck, Innsbruck, Austria Elena Benito-Peña Department of Analytical Chemistry, Complutense University of Madrid, Madrid, Spain Cassiano Augusto Rolim Bernardino Department of Civil Engineering, COPPE, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil Prism Bharadwaj Material Research Application Laboratory, Department of Nano Science and Materials, Central University of Jammu, Samba, India Ashok Bhaskarwar Department of Chemical Engineering, Indian Institute of Technology, Delhi, India Ashlesha Bhide Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States Azra Bihorac Department of Nephrology, University of Florida, Gainesville, FL, United States Yumna Birjis Department of Electrical and Computer Engineering, Faculty of Engineering, University of Windsor, Windsor, ON, Canada

Contributors For All Volumes

Vincent Blay Division of Biomaterials and Bioengineering, University of California San Francisco, San Francisco, CA, United States Michele Borgese Siae Microelettronica, Milano, Italy Alper Bozkurt Department of Electrical and Computer Engineering, NC State University, Raleigh, NC, United States Maria Luisa Braunger Department of Applied Physics, “Gleb Wataghin” Institute of Physics (IFGW), University of Campinas (UNICAMP), Campinas, Brazil; and School of Technology and Sciences (FCT), São Paulo State University (UNESP), Presidente Prudente, Brazil Bernardo Ferreira Braz Analytical Development Laboratory, Institute of Chemistry, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil Kelly Brown WestChem. Department of Pure and Applied Chemistry, Technology and Innovation Centre, University of Strathclyde, Glasgow, United Kingdom Rafael M Buoro Department of Molecular Physics and Chemistry, Institute of Chemistry of São Carlos (IQSC), University of São Paulo, São Carlos, Brazil François Buret Laboratoire Ampère, École Centrale de Lyon, CNRS UMR 5005, Université de Lyon, Écully, France Sebastián Cajigas Max Planck Tandem Group in Nanobioengineering, University of Antioquia, Medellín, Colombia Cinzia Caliendo Istituto di Fotonica e Nanotecnologie, IFN-CNR, Unit of Rome, Rome, Italy Berk Camli Department of Electrical & Electronics Engineering, Bogazici University, Istanbul, Turkey Paulo Victor Soares Campos Instituto de Química - University of Rio de Janeiro, Rio de Janeiro, Brazil Alexandra Canciu Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania

xiii

Giada Caniglia Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany Soledad Cárdenas Affordable and Sustainable Sample Preparation (AS2P) Research Group, Departamento de Química Analítica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUNAN, Universidad de Córdoba, Córdoba, Spain Arnaldo Alves Cardoso Institute of Chemistry, São Paulo State University (UNESP), Araraquara, Brazil Rafael M Cardoso Federal University of Uberlândia, Uberlândia, Minas Gerais, Brazil Marco Carminati Dipartimento di Elettronica, Informazione e Bioingegneria, Politecnico di Milano, Milan, Italy SV Carneiro Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil Enrique Javier Carrasco-Correa Department of Analytical Chemistry, University of Valencia, Valencia, Spain Yunus Celik Department of Computer and Information Sciences, Northumbria University, Newcastle, United Kingdom Aatreya Chakravarti Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA, United States Heena Chandel Department of Biotechnology, School of Basic Sciences, Indian Institute of Information Technology Una, Una, India Theodora Chaspari Texas A&M University, College Station, TX, United States Amita Chaudhary Department of Chemical Engineering, Nirma University, Ahmedabad, India Yin Chen Institute of Microbiology, Zhejiang Provincial Center for Disease Control and Prevention, Hangzhou, China

xiv

Contributors For All Volumes

Zetao Chen Biosensor National Special Laboratory, Key Laboratory for Biomedical Engineering of Education Ministry, Department of Biomedical Engineering, Zhejiang University, Hangzhou, China Hyeon-Yeol Cho Department of Bio & Fermentation Convergence Technology, Kookmin University, Seoul, Republic of Korea Jeong-Woo Choi Department of Chemical & Biomolecular Engineering, Sogang University, Seoul, Republic of Korea Yong-Joon Choi Department of Electrical and Electronic Information Engineering, Toyohashi University of Technology, Toyohashi, Japan Ki H. Chon Department of Biomedical Engineering, University of Connecticut, Storrs, CT, United States DDL Chung Composite Materials Research Laboratory, Department of Mechanical and Aerospace Engineering, University at Buffalo, The State University of New York, Buffalo, NY, United States Fernando Henrique Cincotto Analytical Development Laboratory, Institute of Chemistry, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil; and National Institute of Science & Technology of Bioanalytics (INCTBio), Campinas, Brazil CS Clemente Departamento de Química Orgânica e Inorgânica, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil Şule Çolak Department of Electrical and Electronics Engineering, Adana Alparslan Türkeş Science and Technology University, Adana, Turkey Wendell KT Coltro Instituto de Química, Universidade Federal de Goiás, Goiânia, Brazil; and Instituto Nacional de Ciência e Tecnologia de Bioanalítica e INCTBio, Campinas, Brazil Josiele Aparecida Magalhães Conrado Institute of Chemistry, Federal University of Uberlandia, Uberlândia, Brazil Marco Consales Optoelectronics Group, Department of Engineering, University of Sannio, Benevento, Italy

James Scott Cooper CSIRO, Manufacturing, Sydney, NSW, Australia Daniel S Correa Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, Sao Carlos, Brazil; and PPGQ, Department of Chemistry, Center for Exact Sciences and Technology, Federal University of Sao Carlos (UFSCar), Sao Carlos, Brazil M Bulut Coskun Erik Jonsson School of Engineering and Computer Science, The University of Texas at Dallas, Richardson, TX, United States José Manuel Costa-Fernández Department of Physical and Analytical Chemistry, University of Oviedo, Oviedo, Spain Fernanda P Costa Department of Fundamental Chemistry, Institute of Chemistry (IQUSP), University of São Paulo, São Paulo, Brazil Filippo Costa Department of Information Engineering, University of Pisa, Pisa, Italy VM Costa Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil Ian M Costanzo Worcester Polytechnic Institute, Electrical and Computer Engineering Department, Worcester, MA, United States Graham Coulby Department of Computer and Information Sciences, Northumbria University, Newcastle, United Kingdom Rosa AS Couto LAQV/REQUIMTE, Department of Chemical Sciences, Faculty of Pharmacy, University of Porto, Porto, Portugal Robert D Crapnell Manchester Metropolitan University, Manchester, United Kingdom Cecilia Cristea Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania Emma Crooke CSIRO, Energy, Perth, WA, Australia AAC Cruz Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil

Contributors For All Volumes

Loanda R Cumba National Centre for Sensor Research, School of Chemical Sciences, Dublin City University, Glasnevin, Ireland

Manali Datta Amity Institute of Biotechnology, Amity University Rajasthan, Jaipur, India

Julie Cupka College of Medicine, University of Florida, Gainesville, FL, United States

Van Dau School of Engineering and Built Environment, Griffith University, Nathan, QLD, Australia

Andrea Cusano Optoelectronics Group, Department of Engineering, University of Sannio, Benevento, Italy

Gabriela De Alvarenga Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil

Abhishek Singh Dahiya Bendable Electronics and Sensing Technologies (BEST) Group, University of Glasgow, Glasgow, United Kingdom Ravinder Dahiya Bendable Electronics and Sensing Technologies (BEST) Group, University of Glasgow, Glasgow, United Kingdom Michael A Daniele Department of Electrical & Computer Engineering, North Carolina State University, Raleigh, NC, United States; and Joint Department of Biomedical Engineering, North Carolina State University and University of North Carolina at Chapel Hill, Raleigh, NC, United States Dzung Dao Queensland Micro-and Nanotechnology Centre, Griffith University, Nathan, QLD, Australia; and School of Engineering and Built Environment, Griffith University, Nathan, QLD, Australia María del Mar Darder Department of Organic Chemistry, Faculty of Chemistry, Optical Chemosensors and Applied Photochemistry Group (GSOLFA), Universidad Complutense de Madrid, Madrid, Spain Sajal K Das Department of Computer Science, Missouri S&T, Rolla, MO, United States Iranaldo S da Silva Chemistry Technology Department, Science and Technology Center, Maranhão Federal University, São Luís, Brazil Rafael Luiz da Silva Department of Electrical and Computer Engineering, NC State University, Raleigh, NC, United States João Flávio da Silveira Petruci Institute of Chemistry, Federal University of Uberlandia, Uberlândia, Brazil Anurup Datta UiT The Arctic University of Norway, Tromsø, Norway

xv

Renato E de Araujo Laboratory of Biomedical Optics and Imaging, Federal University of Pernambuco, Recife, Brazil William Reis de Araujo Portable Chemical Sensors Lab, Department of Analytical Chemistry, Institute of Chemistry, State University of Campinas e UNICAMP, Campinas, Brazil Anerise de Barros Institute of Chemistry, University of Campinas, UNICAMP, Campinas, Brazil Cecilia de Carvalho Castro Silva MackGraphe e Graphene and Nanomaterials Research Center, Mackenzie Presbyterian University, São Paulo, Brazil Miren Ruiz De Eguilaz National Centre for Sensor Research, School of Chemical Sciences, Dublin City University, Glasnevin, Ireland María del Carmen Díaz-Liñán Affordable and Sustainable Sample Preparation (AS2P) Research Group, Departamento de Química Analítica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUNAN, Universidad de Córdoba, Campus de Rabanales, Edificio Marie Curie, Córdoba, Spain Lucas Felipe de Lima Portable Chemical Sensors Lab, Department of Analytical Chemistry, Institute of Chemistry, State University of Campinas e UNICAMP, Campinas, Brazil Andrei Deller Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil Ignacio del Villar Department of Electrical, Electronic and Communication Engineering, Institute of Smart Cities, Public University of Navarre, Pamplona, Spain

xvi

Contributors For All Volumes

Christiano JS de Matos MackGrapheGraphene and Nanomaterials Research Center, Mackenzie Presbyterian University, São Paulo, Brazil Hao Deng Department of Mechanical and Aerospace Engineering, Monash University, Clayton, VIC, Australia Adil Denizli Department of Chemistry, Hacettepe University, Ankara, Turkey Lynn Dennany WestChem. Department of Pure and Applied Chemistry, Technology and Innovation Centre, University of Strathclyde, Glasgow, United Kingdom Rafael Furlan de Oliveira Brazilian Nanotechnology National Laboratory (LNNano), Brazilian Center for Research in Energy and Materials, CNPEM, Campinas, Brazil

Girolamo Di Francia ENEA CR-Portici, Portici, Italy Su Ding School of Advanced Materials and Nanotechnology, Xidian University, Xi’an, China Toan Dinh Centre for Future Materials, University of Southern Queensland, Springfield, QLD, Australia; School of Mechanical and Electrical Engineering, University of Southern Queensland, Springfield, QLD, Australia; and Queensland Micro-and Nanotechnology Centre, Griffith University, Nathan, QLD, Australia Bowei Dong Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore; and NUS Suzhou Research Institute (NUSRI), Suzhou, China

Rafael de Oliveira Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil

Feng Dong Tianjin Key Laboratory of Process Measurement and Control, School of Electrical and Information Engineering, Tianjin University, Tianjin, China

Pedro Henrique de Souza Borges Federal University of Uberlândia, Uberlândia, Brazil

Jinhua Dong School of Life Science and Technology, Weifang Medical University, Weifang, China

Vikram Narayanan Dhamu Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States Fernando Díaz-Barriga División de Estudios Superiores para la Equidad, Facultad de Medicina Universidad Autónoma de San Luis Potosí, San Luis Potosí, Mexico; and Center for Applied Research on Environment and Health (CIAAS), San Luis Potosí, Mexico Lorena Díaz de León-Martínez División de Estudios Superiores para la Equidad, Facultad de Medicina Universidad Autónoma de San Luis Potosí, San Luis Potosí, Mexico; and LABINNOVA Inc., Research Center in Breath for Early Diseases Screening, Zapopan, Mexico Francesco Alessio Dicandia IDS Ingegneria dei Sistemi SpA, Pisa, Italy James Dieffenderfer Department of Electrical and Computer Engineering, NC State University, Raleigh, NC, United States

Shaojun Dong State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, China; and University of Science and Technology of China, Hefei, China Rafael M Dornellas Fluminense Federal University, Niteroi, Rio de Janeiro, Brazil Danilo M dos Santos Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, Sao Carlos, Brazil Ana-Maria Dragan Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania Yanyi Duan School of Chemical Engineering and Technology, Tianjin University, Tianjin, China

Contributors For All Volumes

Malikeh P Ebrahim Department of Electrical and Computer Systems Engineering, Monash University, Melbourne, VIC, Australia Hesham El-Sayed College of Information Technology, United Arab Emirates University, Abu Dhabi, United Arab Emirates Arianna Elefante PolySense Lab e Dipartimento Interateneo di Fisica, University and Politecnico of Bari, Bari, Italy Robert Ellis Koneksa Health, Boston, MA, United States Arezoo Emadi Department of Electrical and Computer Engineering, Faculty of Engineering, University of Windsor, Windsor, ON, Canada Sezgin Ersoy Marmara University, Technology Faculty, Mechatronic Engineering Department, Istanbul, Turkey; and Arçelik A.Ş., Istanbul, Turkey Murilo HM Facure Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, Sao Carlos, Brazil; PPGQ, Department of Chemistry, Center for Exact Sciences and Technology, Federal University of Sao Carlos (UFSCar), Sao Carlos, Brazil; and Department of Chemistry, Center for Exact Sciences and Technology, Federal University of Sao Carlos (UFSCar), Sao Carlos, Brazil Fahimeh Mohagheghian Department of Biomedical Engineering, University of Connecticut, Storrs, CT, United States Luigi Falciola Dipartimento di Chimica, Gruppo di Chimica Elettroanalitica (ELAN), Università degli Studi di Milano, Milano, Italy Kinde Anlay Fante Jimma Institute of Technology, Jimma University, Jimma, Ethiopia Sajid Farooq Institute of Technological Innovation, University of Pernambuco, Recife, Brazil LMUD Fechine Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil

xvii

PBA Fechine Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil Bogdan Feier Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania Julia Oliveira Fernandes Analytical Development Laboratory, Institute of Chemistry, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil André Lopes Ferreira Portable Chemical Sensors Lab, Department of Analytical Chemistry, Institute of Chemistry, State University of Campinas e UNICAMP, Campinas, Brazil Marystela Ferreira Center of Science and Technology for Sustainability (CCTS), Federal University of São Carlos, UFSCar, Sorocaba, Brazil Tiago Luiz Ferreira Institute of Environmental, Chemical and Pharmaceutical Sciences, Federal University of São Paulo, Diadema, Brazil Ali Firoozbakhtian Nanobiosensors Lab, Faculty of New Sciences & Technologies, Department of Life Science Engineering, University of Tehran, Tehran, Iran; and Nanobiosensors Lab, Department of Life Science Engineering, Faculty of New Sciences & Technologies, University of Tehran, Tehran, Iran Rogelio Flores-Ramírez Center for Applied Research on Environment and Health (CIAAS), San Luis Potosí, Mexico Anna Fogde Laboratory of Molecular Science and Engineering, Åbo Akademi University, Turku, Finland Ronen Fogel Biotechnology Innovation Centre, Rhodes University, Makhanda, South Africa Daniel Fong Department of Electrical and Computer Engineering, University of California Davis, Davis, CA, United States Jéssica ES Fonsaca MackGrapheGraphene and Nanomaterials Research Center, Mackenzie Presbyterian University, São Paulo, Brazil

xviii

Contributors For All Volumes

Robert J Forster National Centre for Sensor Research, School of Chemical Sciences, Dublin City University, Glasnevin, Ireland

Maria Isabel Gaviria GDCON Research Group, Engineering Faculty, Universidad de Antioquia, Medellin, Colombia

Giancarlo Fortino Department of Informatics, Modeling, Electronic, and Systems Engineering, University of Calabria, Rende, Italy

Asim H Gazi Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States

Laermer Franz Robert Bosch GmbH, Renningen, Germany

Duygu Nazan Gençoglan Department of Electrical and Electronics Engineering, Adana Alparslan Türkeş Science and Technology University, Adana, Turkey

RM Freire Departamento de Física/CEDENNA, Universidad de Santiago de Chile, USACH, Santiago, Chile; and Instituto de Ciências Químicas Aplicadas, Universidad Autónoma de Chile, Santiago, Chile Beatriz Fresco-Cala Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany Jiye Fu State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Hisakage Funabashi Unit of Biotechnology, Division of Biological and Life Sciences, Graduate School of Integrated Sciences for Life, Hiroshima University, Hiroshima, Japan

Simone Genovesi Department of Information Engineering, University of Pisa, Pisa, Italy Abdelghani Ghanam Equipe Analyses Chimiques et Biocapteurs, Laboratoire Génie des Procédés et Environnement, Faculté des Sciences et Techniques, Université Hassan II de Casablanca, Mohammedia, Maroc; and Laboratoire Ampère, École Centrale de Lyon, CNRS UMR 5005, Université de Lyon, Écully, France Sevda Gharehbaghi Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States

Christophe Furger AOP/MH2F-LAAS CNRS, Toulouse, France

Soheil Ghiasi Department of Electrical and Computer Engineering, University of California Davis, Davis, CA, United States

Antra Ganguly Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States

Tonmoy Ghosh Department of Electrical and Computer Engineering, University of Alabama, Tuscaloosa, AL, United States

Venu G Ganti Bioengineering Graduate Program, Georgia Institute of Technology, Atlanta, GA, United States Bin Gao School of automation engineering, University of Electronic Science and Technology of China, Sichuan, China Alejandro Garcia-Miranda Ferrari Manchester Metropolitan University, Manchester, United Kingdom Candela Melendreras García Department of Physical and Analytical Chemistry, University of Oviedo, Oviedo, Spain

Maysam Ghovanloo Bionic Sciences, Atlanta, GA, United States Hemant Ghyvat Technology University of Denmark, Lyngby, Denmark Ambra Giannetti Istituto di Fisica Applicata “N. Carrara”dIFAC del Consiglio Nazionale delle Ricerche, Florence, Italy Marilena Giglio PolySense Lab e Dipartimento Interateneo di Fisica, University and Politecnico of Bari, Bari, Italy Johannes Glöckler Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany

Contributors For All Volumes

Alan Godfrey Department of Computer and Information Sciences, Northumbria University, Newcastle, United Kingdom Josué M Gonçalves Department of Fundamental Chemistry, Institute of Chemistry, University of Sao Paulo, Sao Paulo, Brazil Luís Moreira Gonçalves Departamento de Química Fundamental, Instituto de Química, Universidade de São Paulo (USP), São Paulo, Brazil Kevin Vega Gonzalez Computer and Information Science and Engineering, University of Florida, Gainesville, FL, United States Prosanta Gope Department of Computer Science University of Sheffield, Sheffield, United Kingdom Justyna Grabska Institute of Analytical Chemistry and Radiochemistry, University of Innsbruck, Innsbruck, Austria Daniel Grasseschi SuperNano Laboratório de Química de Coordenação de Superfícies e Nanomateriais, Instituto de Química, Universidade Federai do Rio de Janeiro (UFRJ), Rio de Janeiro, Brazil Raffaele Gravina Department of Informatics, Modeling, Electronic, and Systems Engineering, University of Calabria, Rende, Italy Man Bock Gu Department of Biotechnology, College of Life Sciences and Biotechnology, Korea University, Seoul, Republic of Korea Bárbara GS Guinati Instituto de Química, Universidade Federal de Goiás, Goiânia, Brazil Ulkuhan Guler Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA, United States; and Worcester Polytechnic Institute, Electrical and Computer Engineering Department, Worcester, MA, United States Jinhong Guo School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China

xix

Jiuchuan Guo School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu, China Xinge Guo Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; and Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore Rui JC Gusmão Department of Inorganic Chemistry, University of Chemistry and Technology, Prague, Czech Republic Naoufel Haddour Laboratoire Ampère, École Centrale de Lyon, CNRS UMR 5005, Université de Lyon, Écully, France Hossam Haick Department of Chemical Engineering and Russell Berrie Nanotechnology Institute, Technion-Israel Institute of Technology, Haifa, Israel Azrul Azlan Hamzah Institute of Microengineering and Nanoelectronics (IMEN), Universiti Kebangsaan Malaysia, Bangi, Malaysia Dong Han Department of Biomedical Engineering, University of Connecticut, Storrs, CT, United States Mahammad H Hasan Department of Mechanical Engineering, California Polytechnic State University, San Luis Obispo, CA, United States Tianyiyi He Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore; and NUS Suzhou Research Institute (NUSRI), Suzhou, China Andreas Hellmann Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany José Manuel Herrero-Martínez Department of Analytical Chemistry, University of Valencia, Valencia, Spain Carla Hertleer Department of Materials, Textiles and Chemical Engineering, Ghent University, Gent, Belgium

xx

Contributors For All Volumes

John S Ho Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore Morteza Hosseini Nanobiosensors Lab, Faculty of New Sciences & Technologies, Department of Life Science Engineering, University of Tehran, Tehran, Iran Oana Hosu Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania Bruna M Hryniewicz Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil Renliang Huang School of Marine Science and Technology, Tianjin University, Tianjin, China Christian W Huck Institute of Analytical Chemistry and Radiochemistry, University of Innsbruck, Innsbruck, Austria

Tarikul Islam Department of Electrical Engineering, Faculty of Engineering and Technology, Jamia Millia Islamia (Central University), New Delhi, India Yuzuru Iwasaki NTT Device Technology labs., NTT Corporation, Atsugi, Japan Elena S Izmailova Koneksa Health, Boston, MA, United States Jana Jágerská UiT The Arctic University of Norway, Tromsø, Norway Bruno Campos Janegitz Department of Nature Sciences, Mathematics and Education, Federal University of São Carlos, São Paulo, Brazil Marisol Jaramillo GIBEC Research Group, Life Sciences Faculty, Universidad EIA, Envigado, Colombia Artur Jedrzak Faculty of Chemical Technology, Institute of Chemical Technology and Engineering, Pozna n University of Technology, Pozna n, Poland

Cesar S Huertas Integrated Photonics and Applications Centre (InPAC), School of Engineering, RMIT University, Melbourne, VIC, Australia

Ananthakrishnan Soundaram Jeevarathinam Department of Biomedical Engineering, Texas A&M University, College Station, TX, United States

Tan-Phat Huynh Laboratory of Molecular Science and Engineering, Åbo Akademi University, Turku, Finland

Teofil Jesionowski Faculty of Chemical Technology, Institute of Chemical Technology and Engineering, Pozna n University of Technology, Pozna n, Poland

Henry Alexander Ignatious College of Information Technology, United Arab Emirates University, Abu Dhabi, United Arab Emirates Kenta Iitani Department of Biomedical Devices and Instrumentation, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo, Japan Muhammad A Imran James Watt School of Engineering, University of Glasgow, Glasgow, United Kingdom Omer T Inan Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States

Yuqi Jiang Division of Biomedical Engineering Research, Department of Surgery, The Chinese University of Hong Kong, Shatin, Hong Kong Yifei Jin Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA, United States Hewon Jung Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States Mehrdad Karimzadehkhouei Department of Mechanical Engineering, Koç University, Istanbul, Turkey

Contributors For All Volumes

Ayesha Kausar Nanosciences Division, National Centre For Physics, Quaid-i-Azam University, Islamabad, Pakistan Ryuji Kawano Department of Biotechnology and Life Science, Tokyo University of Agriculture and Technology, Tokyo, Japan Benedikt Keitel Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany Anwar Ulla Khan Department of Electrical Engineering Technology, College of Applied Industrial Technology (CAIT), Jazan University, Jazan, Kingdom of Saudi Arabia Manzoor Ahmed Khan College of Information Technology, United Arab Emirates University, Abu Dhabi, United Arab Emirates Hyonchol Kim Cellular and Molecular Biotechnology Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Ibaraki, Japan; and Graduate School of Engineering, Tokyo University of Agriculture and Technology, Tokyo, Japan Sang Hoon Kim Department of Applied Physical Sciences, University of North Carolina, Chapel Hill, NC, United States Jacob P Kimball Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States Dmitry Kireev Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, United States Yasutaka Kitahama National Institute of Advanced Industrial Science and Technology, Tsukuba, Japan Vanessa Klobukoski Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil Keiichiro Koiwai Department of Biotechnology and Life Science, Tokyo University of Agriculture and Technology, Tokyo, Japan

xxi

Jürgen Kosel Silicon Austria Labs, Villach, Austria Ketan Kotecha Symbiosis Centre for Applied Artificial Intelligence, Symbiosis International (Deemed) University, Pune, India Christine Kranz Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany Mahesh Kumar Department of Electrical Engineering, Indian Institute of Technology, Jodhpur, India Pawan Kumar Material Research Application Laboratory, Department of Nano Science and Materials, Central University of Jammu, Samba, India Vinod Kumar Defence Research and Development Establishment, Gwalior, India Yogeenth Kumaresan Bendable Electronics and Sensing Technologies (BEST) Group, University of Glasgow, Glasgow, United Kingdom Maria Kuznowicz Faculty of Chemical Technology, Institute of Chemical Technology and Engineering, Pozna n University of Technology, Pozna n, Poland Burcu Arman Kuzubaşoglu Faculty of Textile Technologies and Design, Department of Textile Engineering, Istanbul Technical University, Istanbul, Turkey Rafaela Silva Lamarca National Institute for Alternative Technologies for Detection, Toxicological Evaluation and Removal of Micropollutants and Radioactive Materials (INCTDATREM), Institute of Chemistry, São Paulo State University (UNESP), Araraquara, Brazil Guillermo Lasarte-Aragonés Affordable and Sustainable Sample Preparation (AS2P) Research Group, Departamento de Química Analítica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUNAN, Universidad de Córdoba, Córdoba, Spain Tahmid Latif Department of Electrical and Computer Engineering, NC State University, Raleigh, NC, United States

xxii

Contributors For All Volumes

Tomi Laurila Department of Electrical Engineering and Automation, School of Electrical Engineering, Aalto University, Espoo, Finland; and Department of Chemistry and Materials Science, School of Chemical Engineering, Aalto University, Espoo, Finland Bang Hyun Lee Department of Electrical & Computer Engineering, North Carolina State University, Raleigh, NC, United States Chengkuo Lee Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore; NUS Suzhou Research Institute (NUSRI), Suzhou, China; and NUS Graduate School for Integrative Science and Engineering, National University of Singapore, Singapore, Singapore Jinhee Lee The University of North Carolina at Chapel Hill and North Carolina State University, Joint Department of Biomedical Engineering, Chapel Hill, NC, United States Yeng Seng Lee Faculty of Electronic Engineering Technology, Universiti Malaysia Perlis (UniMAP), Perlis, Malaysia; and Advanced Communication Engineering (ACE) Centre of Excellence, Universiti Malaysia Perlis (UniMAP), Perlis, Malaysia Chuanxi Li School of Chemical Engineering and Technology, Tianjin University, Tianjin, China Haoran Li School of automation engineering, University of Electronic Science and Technology of China, Sichuan, China Wanli Li Center of Micro-Nano Engineering, School of Mechanical Engineering, Jiangnan University, Wuxi, China Zheng Li Institute for Advanced Study, Shenzhen University, Shenzhen, China Zhipeng Li Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore

Teik-Cheng Lim School of Science and Technology, Singapore University of Social Sciences, Singapore, Singapore JPO Lima Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil Renato S Lima Brazilian Nanotechnology National Laboratory (LNNano), Brazilian Center for Research in Energy and Materials (CNPEM), Campinas, Brazil; Institute of Chemistry, University of Campinas, Campinas, Brazil; and Center of Natural and Human Sciences, Federal University of ABC, Santo André, Brazil Paulo Clairmont F de Lima Gomes National Institute for Alternative Technologies for Detection, Toxicological Evaluation and Removal of Micropollutants and Radioactive Materials (INCTDATREM), Institute of Chemistry, São Paulo State University (UNESP), Araraquara, Brazil Janice Limson Biotechnology Innovation Centre, Rhodes University, Makhanda, South Africa Jing Liu Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China Qing Liu Department of Electrical and Electronic Engineering, Xi’an Jiaotong-Liverpool University, Suzhou, China Qingjun Liu Biosensor National Special Laboratory, Key Laboratory for Biomedical Engineering of Education Ministry, Department of Biomedical Engineering, Zhejiang University, Hangzhou, China Taoping Liu Interdisciplinary Research Center of Smart Sensors, Academy of Advanced Interdisciplinary Research, Xidian University, Xi’an, China Alnilan Lobato Departamento de Química Fundamental, Instituto de Química, Universidade de São Paulo (USP), São Paulo, Brazil Edgar Lobaton Department of Electrical and Computer Engineering, NC State University, Raleigh, NC, United States

Contributors For All Volumes

xxiii

Tyler Loftus Department of Surgery, University of Florida, Gainesville, FL, United States

Akash Kumar Maity Electrical and Computer Engineering, Rice University, Houston, TX, United States

Ángela I López-Lorente Affordable and Sustainable Sample Preparation (AS2P) Research Group, Departamento de Química Analítica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUNAN, Universidad de Córdoba, Córdoba, Spain

Benny Malengier Department of Materials, Textiles and Chemical Engineering, Ghent University, Ghent, Belgium Giuliano Manara Department of Information Engineering, University of Pisa, Pisa, Italy

Calvin Love Department of Electrical and Computer Engineering, Faculty of Engineering, University of Windsor, Windsor, ON, Canada

Subhadeep Mandal Department of Soil Science and Agricultural Chemistry, Uttar Banga Krishi Viswavidyalaya, Cooch Behar, India

Christopher R Lowe Cambridge Academy of Therapeutic Sciences (CATS), University of Cambridge, Cambridge, United Kingdom

D Manoj Planning and Monitoring Cell, National Institute of Food Technology, Entrepreneurship and Management, Thanjavur, India

Yanli Lu Biosensor National Special Laboratory, Key Laboratory for Biomedical Engineering of Education Ministry, Department of Biomedical Engineering, Zhejiang University, Hangzhou, China

Danilo Manzani Department of Chemistry and Molecular Physics, São Carlos Institute of Chemistry, University of São Paulo, São Carlos, Brazil

Rafael Lucena Affordable and Sustainable Sample Preparation (AS2P) Research Group, Departamento de Química Analítica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUNAN, Universidad de Córdoba, Córdoba, Spain Qiuping Ma School of automation engineering, University of Electronic Science and Technology of China, Sichuan, China Samer Mabrouk Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States Grace M Maddocks Department of Electrical & Computer Engineering, North Carolina State University, Raleigh, NC, United States Mohammad Mahdavi Erik Jonsson School of Engineering and Computer Science, The University of Texas at Dallas, Richardson, TX, United States Claudio Fernando Mahler Department of Civil Engineering, COPPE, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

Lanqun Mao College of Chemistry, Beijing Normal University, Beijing, China Ulm Markus Bosch Sensortec GmbH, Reutlingen, Germany Flávia C Marques Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Hélène Martin-Yken Toulouse Biotechnology Institute, Bio & Chemical Engineering, Université de Toulouse, INSA TBIdINSA Toulouse, Toulouse, France William S Martini Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Paulo R Martins Institute of Chemistry, Federal University of Goias, Goiania, Brazil Ettore Massera ENEA CR-Portici, Portici, Italy

xxiv

Contributors For All Volumes

Ignacio R Matias Department of Electrical, Electronic and Communication Engineering, Institute of Smart Cities, Public University of Navarre, Pamplona, Spain

Daimei Miura Department of Biotechnology and Life Science, Graduate School of Engineering, Tokyo University of Agriculture and Technology, Tokyo, Japan

Avik Mazumder Defence Research and Development Establishment, Gwalior, India

Boris Mizaikoff Institute of Analytical and Bioanalytical Chemistry, Ulm University, Ulm, Germany; and Hahn-Schickard, Institute for Microanalysis Systems, Ulm, Germany

Douglas S McGregor Kansas State University, Mechanical and Nuclear Engineering, Manhattan, KS, United States Michael J McShane Department of Biomedical Engineering, Texas A&M University, College Station, TX, United States; Department of Materials Science and Engineering, Texas A&M University, College Station, TX, United States; and Center for Remote Health Technologies and Systems, Texas Engineering Experiment Station, College Station, TX, United States Luiza A Mercante General and Inorganic Chemistry Department, Institute of Chemistry, Federal University of Bahia (UFBA), Salvador, Brazil Arben Merkoçi Nanobioelectronics & Biosensors Group, Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Bellaterra, Spain; and ICREAdInstitució Catalana de Recerca i Estudis Avançats, Barcelona, Spain Andrea Michel Department of Information Engineering, University of Pisa, Pisa, Italy Ricardo C Michel Instituto de Química, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Maria Lucia Miglietta ENEA CR-Portici, Portici, Italy Fernanda L Migliorini Nanotechnology National Laboratory for Agriculture (LNNA), Embrapa Instrumentação, Sao Carlos, Brazil Kohji Mitsubayashi Department of Biomedical Devices and Instrumentation, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo, Japan

Gita Khalili Moghaddam Department of Clinical Neurosciences, Cambridge Biomedical Campus, Addenbrooke’s Hospital, Cambridge, United Kingdom Ali Mohammadi Department of Electronic and Electrical Engineering, University of Bath, Claverton Down, United Kingdom Hasna Mohammadi Equipe Analyses Chimiques et Biocapteurs, Laboratoire Génie des Procédés et Environnement, Faculté des Sciences et Techniques, Université Hassan II de Casablanca, Mohammedia, Maroc SO Reza Moheimani Erik Jonsson School of Engineering and Computer Science, The University of Texas at Dallas, Richardson, TX, United States Jason Moore Department of Computer and Information Sciences, Northumbria University, Newcastle, United Kingdom Mateus P Moreira SuperNano Laboratório de Química de Coordenação de Superfícies e Nanomateriais, Instituto de Química, Universidade Federai do Rio de Janeiro (UFRJ), Rio de Janeiro, Brazil Nikaele S Moreira Instituto de Química, Universidade Federal de Goiás, Goiânia, Brazil Maria Cruz Moreno-Bondi Department of Analytical Chemistry, Complutense University of Madrid, Madrid, Spain Mohammad Motamedi Department of Electrical and Computer Engineering, University of California Davis, Davis, CA, United States Robert W Motl Department of Kinesiology and Nutrition, College of Applied Health Sciences, University of Illinois Chicago, Chicago, IL, United States

Contributors For All Volumes

xxv

Tamara Dolores Moya-Cavas Department of Analytical Chemistry, Complutense University of Madrid, Madrid, Spain

Karina Nigoghossian ICGM, Univ. Montpellier, CNRS, ENSCM, Montpellier, France

Rodrigo Alejandro Abarza Muñoz Federal University of Uberlândia, Uberlândia, Brazil

Hamed Nikfarjam Department of Electrical Engineering, The University of Texas at Dallas, Richardson, TX, United States

Sriram Muthukumar EnLiSense LLC, Allen, TX, United States Matthew Myers CSIRO, Energy, Perth, WA, Australia Sh Nadzirah Institute of Microengineering and Nanoelectronics (IMEN), Universiti Kebangsaan Malaysia, Bangi, Malaysia Kuniaki Nagamine Graduate School of Organic Materials Science, Yamagata University, Yamagata, Japan; and Research Center of Organic Electronics (ROEL), Yamagata University, Yamagata, Japan Chikashi Nakamura Cellular and Molecular Biotechnology Research Institute, National Institute of Advanced Industrial Science and Technology (AIST), Ibaraki, Japan; and Graduate School of Engineering, Tokyo University of Agriculture and Technology, Tokyo, Japan Nobuhumi Nakamura Department of Biotechnology and Life Science, Tokyo University of Agriculture and Technology, Tokyo, Japan Haleh Nazemi Department of Electrical and Computer Engineering, Faculty of Engineering, University of Windsor, Windsor, ON, Canada Subhash Nerella Biomedical Engineering, University of Florida, Gainesville, FL, United States Tommaso Nespoli Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Politecnico di Milano, Milan, Italy Nam-Trung Nguyen Queensland Micro-and Nanotechnology Centre, Griffith University, Nathan, QLD, Australia Thanh Nguyen Queensland Micro-and Nanotechnology Centre, Griffith University, Nathan, QLD, Australia

Thierry Noguer University Perpignan Via Domitia, BiocapteursAnalyses-Environment, Perpignan, France; and Laboratoire de Biodiversité et Biotechnologies Microbiennes, USR 3579 Sorbonne Universités (UPMC) Paris 6 et CNRS Observatoire Océanologique, Banyuls-sur-Mer, France Edson Nossol Federal University of Uberlândia, Uberlândia, Brazil Mina Nouredanesh Vector Institute, McMaster University, Hamilton, ON, Canada Markellos Ntagios Bendable Electronics and Sensing Technologies (BEST) Group, University of Glasgow, Glasgow, United Kingdom Yoji Okabe Department of Mechanical and Biofunctional Systems, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan Quelle G Olimpio Instituto de Macromoléculas Professora Eloisa Mano, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil Gabriela P Oliveira Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Guilherme P Oliveira Fluminense Federal University, Niteroi, Rio de Janeiro, Brazil JJP Oliveira Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil

xxvi

Contributors For All Volumes

Guillermo Orellana Department of Organic Chemistry, Faculty of Chemistry, Optical Chemosensors and Applied Photochemistry Group (GSOLFA), Universidad Complutense de Madrid, Madrid, Spain Jahir Orozco Max Planck Tandem Group in Nanobioengineering, University of Antioquia, Medellín, Colombia

Camila Pesqueira Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil Hoang-Phuong Phan Queensland Micro-and Nanotechnology Centre, Griffith University, Nathan, QLD, Australia

Yukihiro Ozaki School of Biological and Environmental Sciences, Kwansei Gakuin University, Hyogo, Japan

Paulo HS Picciani Instituto de Macromoléculas Professora Eloisa Mano, Universidade Federal do Rio de Janeiro, Rio de Janeiro, Brazil

Oliver Ozioko Bendable Electronics and Sensing Technologies (BEST) Group, University of Glasgow, Glasgow, United Kingdom

Valentina Pifferi Dipartimento di Chimica, Gruppo di Chimica Elettroanalitica (ELAN), Università degli Studi di Milano, Milan, Italy

Goktug Cihan Ozmen Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States

Filippo Pinelli Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Politecnico di Milano, Milan, Italy

Sharnil Pandya Symbiosis Centre for Applied Artificial Intelligence, Symbiosis International (Deemed) University, Pune, India

Luis Francisco Pinotti Semiconductors, Instruments, and Photonics Department (DSIF), School of Electrical and Computer Engineering (FEEC), University of Campinas (Unicamp), São Paulo, Brazil

Claudio Parolo Nanobioelectronics & Biosensors Group, Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Bellaterra, Spain Leonardo G Paterno Laboratório de Pesquisa em Polímeros e Nanomateriais, Instituto de Química, Universidade de Brasília, Brasília, Brazil Pietro Patimisco PolySense Lab e Dipartimento Interateneo di Fisica, University and Politecnico of Bari, Bari, Italy Anirban Paul Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States Bobby Pejcic CSIRO, Mineral Resources, Perth, WA, Australia Emilia Peltola Department of Mechanical and Materials Engineering, University of Turku, Turku, Finland; and Department of Electrical Engineering and Automation, School of Electrical Engineering, Aalto University, Espoo, Finland

Piyanut Pinyou School of Chemistry, Institute of Science, Suranaree University of Technology, Nakhon Ratchasima, Thailand Bart Plovie Bekegem, Belgium Tiziana Polichetti ENEA CR-Portici, Portici, Italy SMA Pontes Grupo de Química de Materiais Avançados (GQMat), Departamento de Química Analítica e Físico-Química, Universidade Federal do Ceará e UFC, Campus do Pici, Fortaleza, Brazil Carmen CY Poon Division of Biomedical Engineering Research, Department of Surgery, The Chinese University of Hong Kong, Shatin, Hong Kong; and GMed IT Ltd., Shatin, Hong Kong Felix Portillo Department of Electrical and Computer Engineering, University of California Davis, Davis, CA, United States Siavash Pourkamali Department of Electrical Engineering, The University of Texas at Dallas, Richardson, TX, United States

Contributors For All Volumes

Dylan Powell Department of Computer and Information Sciences, Northumbria University, Newcastle, United Kingdom Shalini Prasad Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States Raphael Bacil Prata Institute of Chemistry, University of São Paulo, São Paulo, Brazil Maren S Prediger Gottfried Wilhelm Leibniz University Hanover, Institute of Micro Production Technology, Garbsen, Germany Mihaela Puiu Laboratory for Quality Control and Process Monitoring, University of Bucharest, Bucharest, Romania José Quílez-Alburquerque Department of Organic Chemistry, Faculty of Chemistry, Optical Chemosensors and Applied Photochemistry Group (GSOLFA), Universidad Complutense de Madrid, Madrid, Spain M Beatriz Quinaz LAQV/REQUIMTE, Department of Chemical Sciences, Faculty of Pharmacy, University of Porto, Porto, Portugal Ankit Raghuram Electrical and Computer Engineering, Rice University, Houston, TX, United States Hemangi Ranade Amity Institute of Biotechnology, Amity University Rajasthan, Jaipur, India Sapana Ranwa Department of Electronics and Communication Engineering, National Institute of Technology, Durgapur, India Parisa Rashidi Biomedical Engineering, University of Florida, Gainesville, FL, United States Tomasz Rebis Faculty of Chemical Technology, Institute of Chemistry and Technical Electrochemistry, Pozna n University of Technology, Pozna n, Poland Giulia Regalia Empatica Srl, Milan, Italy Adriana D Rendelucci Department of Fundamental Chemistry, Institute of Chemistry (IQUSP), University of São Paulo, São Paulo, Brazil

xxvii

Daniele Resnati Empatica Srl, Milan, Italy Lawrance Rhein University of Massachusetts Medical School, Department of Pediatrics, Worcester, MA, United States Camilla Machado Gentil Ribeiro Instituto de Química - University of Rio de Janeiro, Rio de Janeiro, Brazil Eduardo Mathias Richter Federal University of Uberlândia, Uberlândia, Brazil Antonio Riul, Jr Department of Applied Physics, “Gleb Wataghin” Institute of Physics (IFGW), University of Campinas (UNICAMP), Campinas, Brazil Diego P Rocha Federal University of Uberlândia, Uberlândia, Minas Gerais, Brazil Diego Pessoa Rocha Federal University of Uberlândia, Uberlândia, Brazil Lucas Carvalho Veloso Rodrigues Department of Fundamental Chemistry, Institute of Chemistry, University of São Paulo, São Paulo, Brazil Guoguang Rong School of Engineering, Westlake University, Hangzhou, China; and Institute of Advanced Study, Westlake Institute for Advanced Study, Hangzhou, China Giulio Rosati Nanobioelectronics & Biosensors Group, Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Bellaterra, Spain Filippo Rossi Department of Chemistry, Materials and Chemical Engineering “Giulio Natta”, Politecnico di Milano, Milan, Italy Nirmalya Roy Department of Information Systems, University of Maryland, Baltimore, MD, United States Satyaki Roy Department of Genetics, University of North Carolina, Chapel Hill, NC, United States Gaige Ru School of automation engineering, University of Electronic Science and Technology of China, Sichuan, China

xxviii

Contributors For All Volumes

Javier Urraca Ruiz Department of Analytical Chemistry, Complutense University of Madrid, Madrid, Spain Matthew Ruppert Health Outcomes and Biomedical Informatics, University of Florida, Gainesville, FL, United States Mahya Saffarpour Department of Electrical and Computer Engineering, University of California Davis, Davis, CA, United States Satoshi Saga Kumamoto University, Kumamoto, Japan Osamu Saito Department of Mechanical and Biofunctional Systems, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan Maiara Oliveira Salles Instituto de Química - University of Rio de Janeiro, Rio de Janeiro, Brazil Angelo Sampaolo PolySense Lab e Dipartimento Interateneo di Fisica, University and Politecnico of Bari, Bari, Italy Ricardo Erthal Santelli Analytical Development Laboratory, Institute of Chemistry, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil; and National Institute of Science & Technology of Bioanalytics (INCTBio), Campinas, Brazil Murilo Santhiago Brazilian Nanotechnology National Laboratory (LNNano), Brazilian Center for Research in Energy and Materials (CNPEM), Campinas, Brazil; and Center of Natural and Human Sciences, Federal University of ABC, Santo André, Brazil Berlane G Santos Department of Fundamental Chemistry, Institute of Chemistry, University of Sao Paulo, Sao Paulo, Brazil Antonio C Sant’Ana Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Caroline G Sanz National Institute of Materials Physics (NIMP), Magurele, Romania Jeffer Eidi Sasaki Department of Sport Sciences, Federal University Triângulo Mineiro, Uberaba, Brazil

Kazuaki Sawada Department of Electrical and Electronic Information Engineering, Toyohashi University of Technology, Toyohashi, Japan Mohamad Sawan School of Engineering, Westlake University, Hangzhou, China; and Institute of Advanced Study, Westlake Institute for Advanced Study, Hangzhou, China Edward Sazonov Department of Electrical and Computer Engineering, University of Alabama, Tuscaloosa, AL, United States Rodrigo Schneider Department of Chemistry, Center for Exact Sciences and Technology, Federal University of Sao Carlos (UFSCar), Sao Carlos, Brazil Tomohito Sekine Graduate School of Organic Materials Science, Yamagata University, Yamagata, Japan; and Research Center of Organic Electronics (ROEL), Yamagata University, Yamagata, Japan Jéssica Soares Guimarães Selva Institute of Chemistry, University of São Paulo, São Paulo, Brazil Thiago Matheus Guimarães Selva Federal Institute of Education, Science and Technology of Rio de Janeiro, Rio de Janeiro, Brazil Felipe S Semaan Fluminense Federal University, Niteroi, Rio de Janeiro, Brazil Devdip Sen Worcester Polytechnic Institute, Electrical and Computer Engineering Department, Worcester, MA, United States Silvia HP Serrano Department of Fundamental Chemistry, Institute of Chemistry (IQUSP), University of São Paulo, São Paulo, Brazil Syed A Shah Faculty Research Centre for Intelligent Healthcare, Coventry University, Coventry, United Kingdom Md Mobashir Hasan Shandhi Department of Electrical and Computer Engineering, Georgia Institute of Technology, Atlanta, GA, United States Chenjing Shang Shenzhen Key Laboratory of Marine Bioresource and Eco-environmental Science, College of Life Sciences and Oceanography, Shenzhen University, Shenzhen, China

Contributors For All Volumes

S Shanmugasundaram Planning and Monitoring Cell and Center of Excellence in Non Thermal Processing, National Institute of Food Technology, Entrepreneurship and Management, Thanjavur, India Lauren Shaw Biotechnology Innovation Centre, Rhodes University, Makhanda, South Africa Qiongfeng Shi Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore; and NUS Suzhou Research Institute (NUSRI), Suzhou, China Yuxing Shi School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu, China Flavio Makoto Shimizu Department of Applied Physics, “Gleb Wataghin” Institute of Physics (IFGW), University of Campinas (UNICAMP), Campinas, Brazil Hiromasa Shimizu Department of Electrical and Electronic Engineering and Department of Applied Physics and Chemical Engineering, Tokyo University of Agriculture and Technology, Koganei, Japan Kan Shoji Department of Mechanical Engineering, Nagaoka University of Technology, Niigata, Japan J Kenneth Shultis Kansas State University, Mechanical and Nuclear Engineering, Manhattan, KS, United States Habdias A Silva-Neto Instituto de Química, Universidade Federal de Goiás, Goiânia, Brazil Samuel Carlos Silva Federal University of Uberlândia, Uberlândia, Brazil Tiago Almeida Silva Department of Chemistry, Federal University of Viçosa, Viçosa, Brazil Neha Singh Department of Information Systems, University of Maryland, Baltimore, MD, United States GR Sinha International Institute of Information Technology (IIIT), Bangalore, India

xxix

Ping Jack Soh Faculty of Electronic Engineering Technology, Universiti Malaysia Perlis (UniMAP), Perlis, Malaysia; Advanced Communication Engineering (ACE) Centre of Excellence, Universiti Malaysia Perlis (UniMAP), Perlis, Malaysia; and ESAT-TELEMIC Research Division, Dept of Electrical Engineering, Katholieke Universiteit Leuven, Leuven, Belgium Neso Sojic University of Bordeaux, Bordeaux INP, ISM, UMR CNRS 5255, Pessac, France; and State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, Jilin, China Ana Soldado Department of Physical and Analytical Chemistry, University of Oviedo, Oviedo, Spain Maria Soler Nanobiosensors and Bioanalytical Applications Group (NanoB2A), Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC, BIST, and CIBERBBN, Barcelona, Spain Mohammadreza Soleymaniha Erik Jonsson School of Engineering and Computer Science, The University of Texas at Dallas, Richardson, TX, United States Laura Soriano-Dotor Affordable and Sustainable Sample Preparation (AS2P) Research Group, Departamento de Química Analítica, Instituto Universitario de Investigación en Química Fina y Nanoquímica IUNAN, Universidad de Córdoba, Córdoba, Spain Dayana Soto Max Planck Tandem Group in Nanobioengineering, University of Antioquia, Medellín, Colombia Lucas R Sousa Instituto de Química, Universidade Federal de Goiás, Goiânia, Brazil Vincenzo Spagnolo PolySense Lab e Dipartimento Interateneo di Fisica, University and Politecnico of Bari, Bari, Italy Jéssica Santos Stefano Department of Nature Sciences, Mathematics and Education, Federal University of São Carlos, São Paulo, Brazil Mathias Strauss Brazilian Nanotechnology National Laboratory (LNNano), Brazilian Center for Research in Energy and Materials (CNPEM), Campinas, Brazil; and Center of

xxx

Contributors For All Volumes

Natural and Human Sciences, Federal University of ABC, Santo André, Brazil Carolina de Medeiros Strunkis Instituto de Química - University of Rio de Janeiro, Rio de Janeiro, Brazil Sam Stuart Department of Sports, Exercise and Rehabilitation, Northumbria University, Newcastle, United Kingdom Rongxin Su School of Marine Science and Technology, Tianjin University, Tianjin, China; and School of Chemical Engineering and Technology, Tianjin University, Tianjin, China Xuyang Sun School of Engineering Medicine, Beihang University, Beijing, China CK Sunil Department of Food Engineering and Center of Excellence for Grain Sciences, National Institute of Food Technology, Entrepreneurship and Management, Thanjavur, India Siddharth Swaminathan Department of Electrical and Computer Engineering, Faculty of Engineering, University of Windsor, Windsor, ON, Canada Sandor Szabo Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania Hasan Riaz Tahir Department of Materials, Textiles and Chemical Engineering, Ghent University, Gent, Belgium Kouta Takeda Institute for Molecular Science, National Institutes of Natural Sciences, Okazaki, Japan Chao Tan Tianjin Key Laboratory of Process Measurement and Control, School of Electrical and Information Engineering, Tianjin University, Tianjin, China Meltem Tekçin Faculty of Textile Technologies and Design, Department of Textile Engineering, Istanbul Technical University, Istanbul, Turkey Chaitra Telang Department of Bioengineering, Biomedical Microdevices and Nanotechnology Laboratory, University of Texas at Dallas, Richardson, TX, United States

Mihaela Tertis Department of Analytical Chemistry, “Iuliu Haţieganu” University of Medicine and Pharmacy, Cluj-Napoca, Romania Simone Tognetti Empatica Srl, Milan, Italy Shizuo Tokito Graduate School of Organic Materials Science, Yamagata University, Yamagata, Japan; and Research Center of Organic Electronics (ROEL), Yamagata University, Yamagata, Japan Paulo HM Toledo Laboratório de Nanoestruturas Plasmônicas, Núcleo de Espectroscopia e Estrutura Molecular, Centro de Estudos de Materiais (CEM), Departamento de Química, Universidade Federal de Juiz de Fora, Juiz de Fora, Brazil Neil Tom Department of Electrical and Computer Systems Engineering, Monash University, Melbourne, VIC, Australia Koji Toma Department of Biomedical Devices and Instrumentation, Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo, Japan Matteo Tonezzer Institute of Materials for Electronics and Magnetism (IMEM), National Research Council (CNR), Trento, Italy Daniele Tosi School of Engineering and Digital Sciences, Nazarbayev University, Nur-Sultan, Kazakhstan; and National Laboratory Astana, Laboratory of Biosensors and Bioinstruments, Nur-Sultan, Kazakhstan Granch Berhe Tseghai Department of Materials, Textiles and Chemical Engineering, Ghent University, Gent, Belgium; and Jimma Institute of Technology, Jimma University, Jimma, Ethiopia Jing Tu State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Tuna B Tufan Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA, United States; and Worcester Polytechnic Institute, Electrical and Computer Engineering Department, Worcester, MA, United States

Contributors For All Volumes

James Tung Centre for BioEngineering and BioTechnology, University of Waterloo, Waterloo, ON, Canada Hiroshi Ueda Laboratory for Chemistry and Life Science, Institute of Innovative Research, Tokyo Institute of Technology, Tokyo, Japan Guy AE Vandenbosch ESAT-TELEMIC Research Division, Dept of Electrical Engineering, Katholieke Universiteit Leuven, Leuven, Belgium Lai Van Duy International Training Institute for Materials Science (ITIMS), Hanoi University of Science and Technology (HUST), Hanoi, Vietnam Lieva Van Langenhove Department of Materials, Textiles and Chemical Engineering, Ghent University, Gent, Belgium Ashok Veeraraghavan Electrical and Computer Engineering, Rice University, Houston, TX, United States María Vergara-Barberán Department of Analytical Chemistry, University of Valencia, Valencia, Spain Madan L Verma Department of Biotechnology, School of Basic Sciences, Indian Institute of Information Technology Una, Una, India Pankratius Victor Bosch Sensortec GmbH, Reutlingen, Germany Marcio Vidotti Grupo de Pesquisa em Macromoléculas e Interfaces, Departamento de Química, Universidade Federal do Paraná, Curitiba, Brazil Rodrigo Vitorio Department of Sports, Exercise and Rehabilitation, Northumbria University, Newcastle, United Kingdom Marek Vlk UiT The Arctic University of Norway, Tromsø, Norway

xxxi

Feng Wen Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore; and NUS Suzhou Research Institute (NUSRI), Suzhou, China Fei Wu Institute of Chemistry, Chinese Academy of Sciences, Beijing, China Linlin Wu State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Weiwei Wu School of Advanced Materials and Nanotechnology, Xidian University, Xi’an, China; and Interdisciplinary Research Center of Smart Sensors, Academy of Advanced Interdisciplinary Research, Xidian University, Xi’an, China Marc Christopher Wurz Gottfried Wilhelm Leibniz University Hanover, Institute of Micro Production Technology, Garbsen, Germany Guohao Xi State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Yinqiang Xia College of Food Science and Engineering, Northwest A&F University, Yangling, Shaanxi, China Jiawen Xie School of Automation Engineering, University of Electronic Science and Technology of China, Chengdu, China Guobao Xu State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, Jilin, China; and University of Science and Technology of China, Hefei, China

Bo Wang School of Behavioural and Health Sciences, Australian Catholic University, North Sydney, NSW, Australia

Meng-Lei Xu State Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University, Changchun, China

Tayyab Waqar Marmara University, Technology Faculty, Mechatronic Engineering Department, Istanbul, Turkey; and Arçelik A.Ş., Istanbul, Turkey

Arda D Yalcinkaya Department of Electrical & Electronics Engineering, Bogazici University, Istanbul, Turkey

xxxii

Contributors For All Volumes

Yang Yang Institute of Deep-Sea Science and Engineering, Chinese Academy of Sciences, Sanya, China

Amir Tofighi Zavareh Department of Biomedical Engineering, Texas A&M University, College Station, TX, United States

Hiroki Yasuga Faculty of Core Research, Ochanomizu University, Tokyo, Japan

Junfeng Zhai State Key Laboratory of Electroanalytical Chemistry, Changchun Institute of Applied Chemistry, Chinese Academy of Sciences, Changchun, China

Huijun Ye School of Chemical Engineering and Technology, Tianjin University, Tianjin, China Keiichi Yoshimatsu Department of Chemistry, Missouri State University, Springfield, MO, United States Kok Yeow You School of Electrical Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, Skudai, Malaysia Fraser Young Department of Computer and Information Sciences, Northumbria University, Newcastle, United Kingdom Fengming Yu Department of Mechanical and Biofunctional Systems, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan Ping Yu Institute of Chemistry, Chinese Academy of Sciences, Beijing, China Ruoxi Yu Division of Biomedical Engineering Research, Department of Surgery, The Chinese University of Hong Kong, Shatin, Hong Kong Mehmet R Yuce Department of Electrical and Computer Systems Engineering, Monash University, Clayton, Melbourne, VIC, Australia Zhao Yue Department of Microelectronics, Nankai University, Tianjin, China Adnan Zahid School of Engineering and Physical Science, Heriot-Watt University; Edinburgh, United Kingdom

Yanjun Zhang Institute of Microbiology, Zhejiang Provincial Center for Disease Control and Prevention, Hangzhou, China Zixuan Zhang Department of Electrical and Computer Engineering, National University of Singapore, Singapore, Singapore; and Center for Intelligent Sensors and MEMS (CISM), National University of Singapore, Singapore, Singapore Bing Zhao State Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University, Changchun, China Qilong Zhao Institute of Biomedical and Health Engineering, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzhen, China Shuang Zhao School of Information Science and Engineering, Yanshan University, Qinhuangdao, China Yongyi Zhao Electrical and Computer Engineering, Rice University, Houston, TX, United States Yuqiao Zheng School of Engineering, Westlake University, Hangzhou, China; and Institute of Advanced Study, Westlake Institute for Advanced Study, Hangzhou, China Peixi Zhu College of Pharmaceutical Sciences, Zhejiang University of Technology, Hangzhou, China Johanna Zikulnig Silicon Austria Labs, Villach, Austria

CONTENTS OF ALL VOLUMES Editor in Chief

v

Section Editors

vii

Contributors For All Volumes

xi

Preface

xliii

VOLUME 1 Sensor Types, edited by Kazunori Ikebukuro Physical Sensors Physical Sensors: Fluorescence Sensors Yong-Joon Choi and Kazuaki Sawada

1

Physical Sensors: Thermal Sensors Toan Dinh, Thanh Nguyen, Hoang-Phuong Phan, Van Dau, Dzung Dao, and Nam-Trung Nguyen

20

Physical Sensors: Optical Sensors Hiromasa Shimizu

34

Physical Sensors: Plasmonic Sensors Yuzuru Iwasaki

49

Physical Sensors: Mechanical Sensors Satoshi Saga

62

Physical Sensors: Acoustic Sensors Osamu Saito, Fengming Yu, and Yoji Okabe

76

Physical Sensors: Magnetic Sensors Marc Christopher Wurz and Maren S Prediger

97

Physical Sensors: Motion Sensors for Physical Activity Assessment Jeffer Eidi Sasaki and Robert W Motl

111

Physical Sensors: Holographic Sensors Christopher R Lowe and Gita Khalili Moghaddam

123

Physical Sensors: Radiation Sensors Douglas S McGregor and J Kenneth Shultis

141

xxxiii

xxxiv

Contents of all Volumes

Chemical Sensors Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors Abdelghani Ghanam, Hasna Mohammadi, Aziz Amine, Naoufel Haddour, and François Buret

161

Chemical Sensors: Impedimetric Electrochemical Sensors Marco Carminati

178

Chemical Sensors: Conductometric Gas Sensors Girolamo Di Francia, Brigida Alfano, Ettore Massera, Maria Lucia Miglietta, and Tiziana Polichetti

189

Chemical Sensors: Acoustic Gas Sensors Jenitha Antony Balasingam, Siddharth Swaminathan, Haleh Nazemi, Calvin Love, Yumna Birjis, and Arezoo Emadi

209

Chemical Sensors: Optical Gas Sensors Zheng Li

226

Chemical Sensors: Photoelectrochemical Sensors Zhao Yue and Shuang Zhao

243

Chemical Sensors: Wearable Sensors Grace M Maddocks and Michael A Daniele

260

Biosensors Biosensors: Enzyme Sensors Kouta Takeda and Nobuhumi Nakamura

281

Biosensors: Immunosensors Daimei Miura and Ryutaro Asano

298

Biosensors: Biosensors Using Engineered Protein Hisakage Funabashi

315

Biosensors: Hybridization-Based Nucleic Acid Sensors Hemangi Ranade and Manali Datta

326

Biosensors: Programmable Nucleic Acid-Binding Protein-Based Nucleic Acid Detection and Biosensing Jinhee Lee

339

Biosensors: Aptamer Sensors Bang Hyun Lee, Sang Hoon Kim, and Man Bock Gu

363

Biosensors: Receptor, Binding Protein, and Peptide Sensors Mihaela Tertis, Alexandra Canciu, Ana-Maria Dragan, Oana Hosu, Sandor Szabo, Bogdan Feier, and Cecilia Cristea

377

Biosensors: Biomimetic Sensors Keiichi Yoshimatsu

393

Biosensors: Microbial Sensors Shaojun Dong and Junfeng Zhai

405

Biosensors: Cell- and Tissue-Containing Biosensors Hélène Martin-Yken and Christophe Furger

420

Biosensors: Biosensors With Signal Amplification Sebastián Cajigas, Dayana Soto, and Jahir Orozco

429

Contents of all Volumes

xxxv

Biosensors: Homogeneous Detection Hiroshi Ueda and Jinhua Dong

458

Biosensors: Gas Sensors Takahiro Arakawa, Kenta Iitani, Koji Toma, and Kohji Mitsubayashi

478

Combinatorial Sensors: An Integrated Approach to Lifestyle Management and Environmental Surveillance Vikram Narayanan Dhamu, Ivneet Banga, Anirban Paul, Antra Ganguly, Ashlesha Bhide, Chaitra Telang, Sriram Muthukumar, and Shalini Prasad

505

New Sensing Technologies New Sensing Technologies: Microtas/NEMS/MEMS Hiroki Yasuga, Kan Shoji, Keiichiro Koiwai, and Ryuji Kawano

526

New Sensing Technologies: Sensors for In Vivo Analysis Fei Wu, Ping Yu, and Lanqun Mao

541

New Sensing Technologies: Atomic Force Microscopy Chikashi Nakamura and Hyonchol Kim

556

New Sensing Technologies: Biosensors Based on Magnetic Nanoparticles and Magnetic Force Microscopy Jeong-Woo Choi and Hyeon-Yeol Cho New Sensing Technologies: Nanopore Sensing Jiye Fu, Linlin Wu, Guohao Xi, and Jing Tu

572 581

VOLUME 2 Materials and Devices, edited by Thiago R L C Paixão Sensing Materials Sensing Materials: Ceramics Sapana Ranwa and Mahesh Kumar

1

Sensing Materials: Glass Tiago Luiz Ferreira

14

Sensing Materials: Carbon Materials Alejandro Garcia-Miranda Ferrari, Robert D Crapnell, and Craig E Banks

25

Sensing Materials: Diamond-Based Materials Thiago Matheus Guimarães Selva, Jéssica Soares Guimarães Selva, and Raphael Bacil Prata

45

Sensing Materials: Electrochemical Sensors Enabled by 3D Printing Diego P Rocha, Renata BA Albuquerque, Guilherme P Oliveira, Rafael M Cardoso, Felipe S Semaan, Rafael M Dornellas, Eduardo M Richter, and Rodrigo AA Muñoz

73

Sensing Materials: Self-Healing Hydrogels Anna Fogde and Tan-Phat Huynh

89

Sensing Materials: Metal Oxides Josué M Gonçalves, Paulo R Martins, Berlane G Santos, Koiti Araki, and Lúcio Angnes

98

Sensing Materials: Liquid Metal-Enabled Flexible Sensors for Biomedical Applications Xuyang Sun and Jing Liu

114

xxxvi

Contents of all Volumes

Sensing Materials: Organic Polymers Paulo HS Picciani, Flavio Makoto Shimizu, Quelle G Olimpio, and Ricardo C Michel

130

Sensing Materials: Hydrogels Filippo Pinelli, Tommaso Nespoli, and Filippo Rossi

148

Sensing Materials: Electrolyte-Gated Organic Field-Effect Transistors (EGOFETs) Cecilia de Carvalho Castro Silva and Luis Francisco Pinotti

167

Sensing Materials: Composites and Hybrid Materials Teik-Cheng Lim

187

Sensing Materials: Self-Sensing Materials DDL Chung

196

Sensing Materials: Bimetallics and Metal Mixtures (Core-Shell Microspheres) Amita Chaudhary, Baijnath, Prism Bharadwaj, Pawan Kumar, and Ashok Bhaskarwar

204

Sensing Materials: Nanomaterials Tiago Almeida Silva, Jéssica Santos Stefano, and Bruno Campos Janegitz

212

Sensing Materials: Electronic Nose and Tongue Materials Camilla Machado Gentil Ribeiro, Carolina de Medeiros Strunkis, Paulo Victor Soares Campos, and Maiara Oliveira Salles

231

Sensing Materials: Functionalized Advanced Carbon-Based Nanomaterials Anerise de Barros, Maria Luisa Braunger, Rafael Furlan de Oliveira, and Marystela Ferreira

254

Sensing Materials: Nanostructured Platforms Based on Conducting Polymers for Sensing Bruna M Hryniewicz, Gabriela De Alvarenga, Andrei Deller, Larissa Bach-Toledo, Camila Pesqueira, Vanessa Klobukoski, and Marcio Vidotti

269

Sensing Materials: Biopolymeric Nanostructures Teofil Jesionowski, Maria Kuznowicz, Artur Je˛ drzak, and Tomasz Re˛ bis

286

Sensing Materials: Nanocomposites Ayesha Kausar

305

Surface Plasmon Resonance Platforms for Chemical and Bio-Sensing Jéssica ES Fonsaca, Mateus P Moreira, Sajid Farooq, Renato E de Araujo, Christiano JS de Matos, and Daniel Grasseschi

316

Sensing Materials: Nanostructured Biomaterials R Abdel-Karim

354

Sensing Materials: Graphene Edson Nossol, Rodrigo Alejandro Abarza Muñoz, Eduardo Mathias Richter, Pedro Henrique de Souza Borges, Samuel Carlos Silva, and Diego Pessoa Rocha

367

(Bio)Sensing Materials: Quantum Dots Julia Oliveira Fernandes, Cassiano Augusto Rolim Bernardino, Bernardo Ferreira Braz, Claudio Fernando Mahler, Ricardo Erthal Santelli, and Fernando Henrique Cincotto

389

Sensing Materials: Electropolymerized Molecularly Imprinted Polymers Rosa AS Couto, Alnilan Lobato, M Beatriz Quinaz, and Luís Moreira Gonçalves

401

Sensing Materials: Enzymes and Aptamers Piyanut Pinyou, Thierry Noguer, and Vincent Blay

413

Sensing Materials: Bio-inspired Materials Qilong Zhao

435

Contents of all Volumes

xxxvii

Sensing Materials: Electrochemical Applications of DNA Sensors and Biosensors Caroline G Sanz, Rafael M Buoro, Raphael P Bacil, Iranaldo S da Silva, Adriana D Rendelucci, Fernanda P Costa, and Silvia HP Serrano

445

Sensing Materials: Lanthanide Materials Lucas Carvalho Veloso Rodrigues, Danilo Manzani, and Karina Nigoghossian

468

Sensing Materials: Novel Electrochemical and Fluorescent Materials Lynn Dennany

483

Sensing Materials: 2D Semiconductors for Biosensing Rui JC Gusmão

505

Sensing Materials: Nanofibers Produced by Electrospinning and Solution Blow Spinning Rafaela S Andre, Murilo HM Facure, Rodrigo Schneider, Fernanda L Migliorini, Danilo M dos Santos, Luiza A Mercante, and Daniel S Correa

521

Sensing Materials: Optical Sensing Based on Carbon Quantum Dots AAC Cruz, SV Carneiro, SMA Pontes, JJP Oliveira, JPO Lima, VM Costa, LMUD Fechine, CS Clemente, RM Freire, and PBA Fechine

542

Sensing Materials: UV/Vis-Based Optical Sensors for Gaseous and Volatile Analytes Diandra Nunes Barreto, Josiele Aparecida Magalhães Conrado, Rafaela Silva Lamarca, Alex Domingues Batista, Arnaldo Alves Cardoso, Paulo Clairmont F de Lima Gomes, and João Flávio da Silveira Petruci

560

Sensing Materials: Paper Substrates Lucas R Sousa, Habdias A Silva-Neto, Nikaele S Moreira, Bárbara GS Guinati, and Wendell KT Coltro

577

Sensing Materials: Flexible Carbon-Based Electrochemical Devices Based on the Three-Dimensional Architecture of Paper Renato S Lima, Mathias Strauss, and Murilo Santhiago

600

Sensing Interfaces Sensing Interfaces: Self-Cleaning Materials for Electroanalytical Applications Wafa Aidli, Valentina Pifferi, and Luigi Falciola

613

Sensing Interfaces: Antifouling Materials for Sensors Rongxin Su, Yinqiang Xia, Chuanxi Li, Huijun Ye, Yanyi Duan, and Renliang Huang

619

Sensing Interfaces: Materials for Wearable Sensors Lucas Felipe de Lima, André Lopes Ferreira, and William Reis de Araujo

636

VOLUME 3 Sensor Development, edited by Mehmet R. Yuce Wearable Sensors and Deep Learning for the Management of Acute Pancreatitis in Precision Medicine Qing Liu, Yuqi Jiang, Ruoxi Yu, and Carmen CY Poon

1

COVID-19 Diagnostic Methods and Detection Techniques Guoguang Rong, Yuqiao Zheng, Yin Chen, Yanjun Zhang, Peixi Zhu, and Mohamad Sawan

17

Fabrication Technologies for Flexible Printed Sensors Johanna Zikulnig and Jürgen Kosel

33

xxxviii

Contents of all Volumes

Soft Sensors for Electronic Skin Abhishek Singh Dahiya, Yogeenth Kumaresan, Oliver Ozioko, Markellos Ntagios, and Ravinder Dahiya

51

Printed Electronics-Enabled Wearable/Portable Physical and Chemical Sensors for Personal Digital Healthcare Kuniaki Nagamine, Tomohito Sekine, and Shizuo Tokito

68

Electronic Textiles (E-Textiles): Fabric Sensors and Material-Integrated Wearable Intelligent Systems Burcu Arman Kuzubas¸oglu, Meltem Tekçin, and Senem Kurs¸un Bahadır

80

Electromagnetic Multiphysics Sensing Nondestructive Testing Bin Gao, Gaige Ru, Qiuping Ma, and Haoran Li

101

Surface Acoustic Wave Sensing Sezgin Ersoy and Tayyab Waqar

129

Fiber Optic Biosensors Daniele Tosi

142

Electrochemical Sensors for New Challenges Lynn Dennany and Kelly Brown

158

Interfacing Circuits for Capacitive Sensors Anwar Ulla Khan and Tarikul Islam

174

Gas Sensors Matteo Tonezzer and Lai Van Duy

185

Biosensor Development Azrul Azlan Hamzah and Sh Nadzirah

209

Passive Sensors for Detection of Food Intake Tonmoy Ghosh and Edward Sazonov

218

Triboelectric Sensors for IoT and Wearable Applications Zixuan Zhang, Xinge Guo, Feng Wen, Qiongfeng Shi, Tianyiyi He, Bowei Dong, and Chengkuo Lee

235

Soft and Stretchable Electronics Design Yang Yang, Su Ding, Bart Plovie, Wanli Li, and Chenjing Shang

258

Radar and Non-Contact Sensing Malikeh P Ebrahim, Neil Tom, Duygu Nazan Gençog lan, S¸ule Çolak, and Mehmet R Yuce

287

Resonant Type RF Glucose Biosensors Berk Camli and Arda D Yalcinkaya

308

Environmental Sensors GR Sinha

332

Vibration-Based Energy Harvesting for Sensors Ali Mohammadi

345

Aptamers as Versatile Tools for Expanding the Scope of Sensors Ronen Fogel, Lauren Shaw, and Janice Limson

352

Radio Frequency Identification (RFID) for Sensing Filippo Costa, Simone Genovesi, Michele Borgese, Andrea Michel, Francesco Alessio Dicandia, and Giuliano Manara

375

Contents of all Volumes

xxxix

Microwave Metamaterials for Biomedical Sensing John S Ho and Zhipeng Li

391

Antennas for Sensing Applications Yeng Seng Lee, Kok Yeow You, Ping Jack Soh, and Guy AE Vandenbosch

402

Sensors for Neonatal Monitoring Ulkuhan Guler, Devdip Sen, Ian M Costanzo, Tuna B Tufan, and Lawrance Rhein

423

Implantable and Wearable Sensors for Assistive Technologies Ulkuhan Guler, Tuna B Tufan, Aatreya Chakravarti, Yifei Jin, and Maysam Ghovanloo

449

Microfluidic Devices for Biosensing Hao Deng and Tuncay Alan

474

Microwave and Terahertz Sensing Qammer H Abbasi, Hasan T Abbas, Syed A Shah, Adnan Zahid, Muhammad A Imran, and Akram Alomainy

489

Linear Regression and Artificial Neural Network (ANN)-based Approaches to Predict Air Pollution Sharnil Pandya, Hemant Ghyvat, Ketan Kotecha, and Prosanta Gope

497

Design, Fabrication, and Characterization of Bio-MEMS Jiawen Xie, Yuxing Shi, Jiuchuan Guo, and Jinhong Guo

512

Sensor Instrumentation for Flow Measurement Chao Tan and Feng Dong

536

Piezoresistive Nanowire-Based Electromechanical Sensors B Erdem Alaca and Mehrdad Karimzadehkhouei

555

Electrical Impedance-Based Electronic Tongues Murilo HM Facure, Maria L Braunger, Luiza A Mercante, Leonardo G Paterno, Antonio Riul, Jr, and Daniel S Correa

567

Acoustic Wave Conductometric Sensors Cinzia Caliendo

591

Active Microcantilevers for Dynamic Mode Atomic Force Microscopy M Bulut Coskun, Mohammadreza Soleymaniha, Mohammad Mahdavi, and SO Reza Moheimani

617

VOLUME 4 Integration, edited by Roozbeh Jafari Sensors on the Wrist Giulia Regalia, Daniele Resnati, and Simone Tognetti

1

Smart Textiles Granch Berhe Tseghai, Hasan Riaz Tahir, Benny Malengier, Carla Hertleer, Kinde Anlay Fante, and Lieva Van Langenhove

21

Sensor Technology for Autonomous Vehicles Henry Alexander Ignatious, Hesham-El-Sayed, and Manzoor Ahmed Khan

35

Sensors in Hospitals Subhash Nerella, Kevin Vega Gonzalez, Julie Cupka, Matthew Ruppert, Tyler Loftus, Azra Bihorac, and Parisa Rashidi

52

xl

Contents of all Volumes

Wearable Cyberphysical Systems for Biomedicine Tahmid Latif, James Dieffenderfer, Rafael Luiz da Silva, Edgar Lobaton, and Alper Bozkurt

63

Bio-Inspired Design of Biosensor Networks Satyaki Roy and Sajal K Das

86

Electronic Tattoos Dmitry Kireev and Deji Akinwande

103

Implantable Sensors Michael J McShane, Amir Tofighi Zavareh, and Ananthakrishnan Soundaram Jeevarathinam

115

Measuring Physiological Parameters Under the Skin Using Visible/NIR Light Ankit Raghuram, Yongyi Zhao, Akash Kumar Maity, and Ashok Veeraraghavan

133

Hardware and Algorithmic Approaches to Combat Motion Artifacts in Photoplethysmographic Data Dong Han, Fahimeh Mohagheghian, and Ki H. Chon Sensors as Neural Computing Units Fadi Alsaleem, Mahammad H Hasan, Hamed Nikfarjam, and Siavash Pourkamali Computational Aspects in Body Sensor Networks (BSNs): From Operating Systems to Data Fusion Raffaele Gravina and Giancarlo Fortino

143 154

173

Transfer Machine Learning Algorithms Neha Singh and Nirmalya Roy

186

Convolutional Neural Network (CNN) Synthesis for Resource-Constrained Platforms Mohammad Motamedi, Felix Portillo, Mahya Saffarpour, Daniel Fong, and Soheil Ghiasi

204

Low-Cost Computational Models for Biomedical Sensors Yanli Lu, Zetao Chen, and Qingjun Liu

223

Noninvasive Multimodal Physiological Sensing Systems Jacob P Kimball, Asim H Gazi, Goktug Cihan Ozmen, Hewon Jung, Md Mobashir Hasan Shandhi, Samer Mabrouk, Sevda Gharehbaghi, Venu G Ganti, and Omer T Inan

236

Sensor Integration for Behavior Monitoring Theodora Chaspari

253

Sensor Integration for Gait Analysis Yunus Celik, Rodrigo Vitorio, Dylan Powell, Jason Moore, Fraser Young, Graham Coulby, James Tung, Mina Nouredanesh, Robert Ellis, Elena S Izmailova, Sam Stuart, and Alan Godfrey

263

Applications, edited by Boris Mizaikoff and Beatriz Fresco Cala Surface-Enhanced Raman Scattering Sensing of Food Contaminants Gustavo FS Andrade, Rafael de Oliveira, Flávia C Marques, William S Martini, Gabriela P Oliveira, Antonio C Sant’Ana, and Paulo HM Toledo

284

Optical Sensors in Medical Diagnosis José Manuel Costa-Fernández, Candela Melendreras García, and Ana Soldado

297

Electrochemiluminescence Sensors in Bioanalysis Ali Firoozbakhtian, Neso Sojic, Guobao Xu, and Morteza Hosseini

317

Contents of all Volumes

xli

Chemiluminescence Sensors in Bioanalysis Ali Firoozbakhtian and Morteza Hosseini

341

Recent Advances in Plasmonic Biosensors for the Detection of Food Allergens Semra Akgönüllü and Adil Denizli

357

Intracellular Biosensing Ambra Giannetti, Barbara Adinolfi, and Francesco Baldini

372

Label-Free Plasmonic Biosensors in Clinical Diagnostics Maria Soler and Cesar S Huertas

383

Plasmonic Biosensors for Medical Applications Mihaela Puiu and Camelia Bala

402

Volatile Organic Compound Exhaled Breath Sensing Lorena Díaz de León-Martínez, Johannes Glöckler, Boris Mizaikoff, Rogelio Flores-Ramírez, and Fernando Díaz-Barriga

421

Surface-enhanced Raman scattering (SERS) Sensing of Biomedicine and Biomolecules Yasutaka Kitahama, Bing Zhao, and Yukihiro Ozaki

441

Surface-enhanced Raman scattering (SERS) Sensors for Food Safety Meng-Lei Xu, Bing Zhao, and Yukihiro Ozaki

456

Plasmonic Biosensors for Food Safety D Manoj, S Shanmugasundaram, and CK Sunil

471

Near-Infrared (NIR) Sensors for Environmental Analysis Krzysztof B Bec, Justyna Grabska, and Christian W Huck

484

Lab on Fiber Technology Towards Advanced and Multifunctional Point-of-Care Platforms for Precision Medicine Marco Consales, Ignacio del Villar, Ignacio R Matias, and Andrea Cusano

504

Optical Biosensors for Environmental Analysis Maria Isabel Gaviria, Juan Pablo Arango, Kaory Barrientos, and Marisol Jaramillo

528

Flurescence Sensors for the Food Industry Guillermo Lasarte-Aragonés, Laura Soriano-Dotor, Ángela I López-Lorente, Rafael Lucena, and Soledad Cárdenas

549

Molecularly Imprinted Polymer-Based Biomimetic Sensors for Food Analysis Maria Cruz Moreno-Bondi, Elena Benito-Peña, Tamara Dolores Moya-Cavas, and Javier Urraca Ruiz

568

Luminescence-Based Sensors for Water Quality Analysis Guillermo Orellana, María del Mar Darder, and José Quílez-Alburquerque

599

Why Sensors Need Microfluidics: Real-World Applications María Vergara-Barberán, Enrique Javier Carrasco-Correa, and José Manuel Herrero-Martínez

614

Miniaturized Electrochemical Biosensors Andreas Hellmann, Giada Caniglia, and Christine Kranz

636

Chem/Bio Sensors for Marine Applications Bobby Pejcic, Matthew Myers, Emma Crooke, and James Scott Cooper

650

Detection of Chemical Warfare Agents With Chemical Sensors Vinod Kumar and Avik Mazumder

667

xlii

Contents of all Volumes

Surface-enhanced Raman spectroscopy (SERS) Sensors for Clinical Analysis María del Carmen Díaz-Liñán, Rafael Lucena, Soledad Cárdenas, and Ángela I López-Lorente

693

Biosensors for Precision Agriculture Subhadeep Mandal and Ganesh Chandra Banik

709

Electronic Nose Sensors for Healthcare Weiwei Wu, Taoping Liu, and Hossam Haick

728

Nanomaterial-Based Fluorescent Biosensors for Monitoring Environmental Pollutants Heena Chandel, Bo Wang, and Madan L Verma

742

Translational Sensor Technology: From the Laboratory to Industry Laermer Franz, Ulm Markus, and Pankratius Victor

755

Point-of-Care Sensors in Clinical Environments Georgina Alarcón-Angeles, Giaan Arturo Álvarez-Romero, Giulio Rosati, Claudio Parolo, and Arben Merkoçi

772

Quartz-Enhanced Photoacoustic Sensors for Environmental Monitoring Marilena Giglio, Arianna Elefante, Pietro Patimisco, Angelo Sampaolo, and Vincenzo Spagnolo

789

Electrochemical Biosensing of Bacteria and Biofilms Robert J Forster, Miren Ruiz De Eguilaz, Ahmed Barhoum, and Loanda R Cumba

800

Waveguide-Based Devices for Infrared and Raman Spectroscopy Jana Jágerská, Sebastián Alberti, Anurup Datta, and Marek Vlk

814

Carbonaceous Nanomaterials for Electrochemical Biosensing Emilia Peltola and Tomi Laurila

839

Molecularly Imprinted Polymer Sensors for Environmental Analysis Benedikt Keitel, Alex D Batista, Boris Mizaikoff, and Beatriz Fresco-Cala

851

Index

869

PREFACE Sensors and biosensors are rapidly finding use in many aspects of everyday life. The US National Cancer Institute defines a sensor as “a device that responds to a stimulus, such as heat, light, or pressure, and generates a signal that can be measured or interpreted” (Cancer). The US National Research Council has defined a biosensor as a “detection device that incorporates (a) a living organism or product derived from living systems (e.g., an enzyme or an antibody) and (b) a transducer to provide an indication, signal, or other form of recognition of the presence of a specific substance in the environment” (Luong et al., 2008). Nanosensors, which have at least one sensor dimension below 100 nm, offer greater sensitivity than conventional microstructured sensors due to their greater surface area to volume ratio; moreover, these devices should be associated with lower manufacturing costs and lower power consumption levels due to their smaller dimensions (AbdelKarim et al., 2020). New types of sensors and sensor data analysis methods are being developed for use in large networks of connected devices; this approach is commonly called the Internet of Things (Laghari et al., 2021). It is anticipated that the Internet of Things approach will enable improvements to transportation, healthcare, agriculture, and many other human endeavors. The development of sensing technology can be traced to activities by researchers in the mid-nineteenth century. For example, Charles Wheatstone popularized a bridge circuit to measure resistance in 1843, and Lord Kelvin demonstrated that iron and copper wires show changes to their electrical resistance on the application of mechanical strain in 1856; these concepts continue to be useful in the development of modern sensors (Jones, 1998; Greenslade, 2017; Rolnick, 1930; Barlian et al., 2009). The development of biosensing technology can be traced to the first demonstration of a glucose sensor containing an oxygen electrode and a membrane containing the enzyme glucose oxidase by Clark and Lyons in 1962 (Clark and Lyons, 1962; Newman and Setford, 2006). The development of new technologies for blood glucose sensing remains the most important activity in the biosensing field due to the large number of diabetes patients and the association between good control over blood glucose levels and a reduced risk of vascular complications in diabetes patients (The Diabetes Control and Complications Trial Research Group, 1993). Over the coming years, the development of wearable sensors combined with Internet of Things approaches is anticipated to facilitate improved patient care outside of healthcare facilities (Miller et al., 2012; Kukkar et al., 2022). In addition, biosensor technologies have become more sophisticated in detecting molecules via minimally invasive or noninvasive methods, including in interstitial fluid saliva, sweat, and tears (Miller et al., 2012; Kukkar et al., 2022). Another focus area in biosensor research is the development of sensors for use at the point of care; the US National Institutes of Health recently developed centers to advance biosensor technology for use in heart disease treatment and HIV/AIDS patient care in resource-limited environments (nibib). There is also a growing focus on the development of sensor technologies for use in detecting changes in air quality and water quality in order to better understand changes associated with pollution as well as environmental phenomena (Johnson et al., 2007; Abraham and Li, 2016). Improved food production and food safety sensor technologies are anticipated to support increased efficiency and reduced contamination levels, respectively (Wang et al., 2006; Rodrigues et al., 2017; Baruah and Dutta, 2009). Understanding the environmental impact of sensor technologies, including the degradation processes and disposal issues associated with sensors, will likely drive the assessment of new sensor technologies over the coming years (Ishii et al., 1994). The goal of the Encyclopedia of Sensors and Biosensors is to provide foundational level information on the materials and technologies that underlie sensors and biosensors. The encyclopedia considers approaches for sensor design, sensor manufacturing, and measurement of sensor performance for sensors that are used in healthcare, environmental engineering, food safety, and many other fields. Kazunori Ikebukuro at Tokyo

xliii

xliv

Preface

University of Agriculture and Technology edited the section of the encyclopedia that considers various types of physical sensors, chemical sensors, and biosensors; in addition, new sensing technologies such as nanopore sensing and nanoelectromechanical system-based sensors are reviewed. Thiago R.L.C. Paixão at the University of São Paulo edited the section on sensing materials, which includes contributions on graphene, quantum dots, and self-cleaning materials. Mehmet Yuce at Monash University was the editor of the sensor development section, which considers the various approaches for preparing sensors for environmental and medical applications. Roozbeh Jafari at Texas A&M University edited the sensor integration section, which includes contributions on the many types of computational approaches for analyzing sensor information. Boris Mizaikoff at the University of Ulm and Beatriz María Fresco Cala at the University of Córdoba edited the sensor applications section, which considers the use of sensors for medical care, agriculture, food safety, and environmental monitoring. I would like to express my sincere appreciation to the section editors and authors, without whom this encyclopedia would not have been possible. I would also like to thank Katarzyna Miklaszewska, Blerina Osmanaj, Sam Crowe, and Oliver Walter at Elsevier for their outstanding work and tireless effort over the past 3 years to bring the encyclopedia to publication. I hope that this work will serve as a useful resource for chemists, engineers, biological scientists, medical clinicians, and others who are involved in advancing the global sensor and biosensor community. Roger J Narayan, UNC/NCSU Joint Department of Biomedical Engineering References Abdel-Karim, R., Reda, Y., Abdel-Fattah, A., 2020. ReviewdNanostructured materials-based nanosensors. J. Electrochem. Soc. 167, 037554. Abraham, S., Li, X., 2016. Design of a low-cost wireless indoor air quality sensor network system. Int. J. Wireless Inf. Networks 23, 57–65. Barlian, A.A., Park, W.T., Mallon, J.R., Rastegar, A.J., Pruitt, B.L., 2009. Semiconductor piezoresistance for microsystems. Proc. IEEE 97 (3), 513–552. Baruah, S., Dutta, J., 2009 Sep. Nanotechnology applications in pollution sensing and degradation in agriculture: a review. Environ. Chem. Lett. 7 (3), 191–204. https://www.cancer.gov/publications/dictionaries/cancer-terms/def/sensor. Clark Jr., L.C., Lyons, C., 1962. Electrode systems for continuous monitoring in cardiovascular surgery. Ann. N. Y. Acad. Sci. 102, 29–45. Greenslade Jr., T.B., 2017 Feb. The scientific legacy of Charles Wheatstone. The Physics Teacher 55, 80–82. Ishii, K., Eubanks, C.F., Di Marco, P., 1994 Jan 1. Design for product retirement and material life-cycle. Materials & Design 15 (4), 225–233. Johnson, K.S., Needoba, J.A., Riser, S.C., Showers, W.J., 2007 Feb 14. Chemical sensor networks for the aquatic environment. Chemical Reviews 107, 623–640. Jones, W.D., 1998. Charles Wheatstone: genius remembered. IEEE Spectrum 35, 68T4. Kukkar, D., Zhang, D., Jeon, B.H., Kim, K.H., 2022. Recent advances in wearable biosensors for non-invasive monitoring of specific metabolites and electrolytes associated with chronic kidney disease: Performance evaluation and future challenges. TrAC Trends in Analytical Chemistry 150, 116570. Laghari, A.A., Wu, K., Laghari, R.A., Ali, M., Khan, A.A., 2021. A review and state of art of Internet of Things (IoT). Archives of Computational Methods in Engineering 14, 1–9. Luong, J.H., Male, K.B., Glennon, J.D., 2008 Sep 1. Biosensor technology: technology push versus market pull. Biotechnology Advances 26, 492–500. Miller, P.R., Skoog, S.A., Edwards, T.L., Lopez, D.M., Wheeler, D.R., Arango, D.C., Xiao, X., Brozik, S.M., Wang, J., Polsky, R., Narayan, R.J., 2012. Multiplexed microneedle-based biosensor array for characterization of metabolic acidosis. Talanta 88, 739–742. Newman, J.D., Setford, S.J., 2006. Enzymatic biosensors. Molecular Biotechnol. 32, 249–268. https://www.nibib.nih.gov/research-program/point-care-technologies-research-network. Rodrigues, S.M., Demokritou, P., Dokoozlian, N., Hendren, C.O., Karn, B., Mauter, M.S., Sadik, O.A., Safarpour, M., Unrine, J.M., Viers, J., Welle, P., 2017. Nanotechnology for sustainable food production: promising opportunities and scientific challenges. Environ. Sci. 4 (4), 767–781. Rolnick, H., 1930 Aug 1. Tension coefficient of resistance of metals. Phys. Rev. 36, 506. The Diabetes Control and Complications Trial Research Group, 1993. The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. New Engl. J. Med. 329, 683–689. Wang, N., Zhang, N., Wang, M., 2006 Jan 1. Wireless sensors in agriculture and food industrydrecent development and future perspective. Comput. Electron. Agric. 50 (1), 1–4.

Physical Sensors: Fluorescence Sensors Yong-Joon Choi and Kazuaki Sawada, Department of Electrical and Electronic Information Engineering, Toyohashi University of Technology, Toyohashi, Japan Published by Elsevier Ltd.

Introduction Fluorescence Principle Fluorescence intensity Time-resolved fluorescence lifetime Fluorescence polarization Fluorescence resonance energy transfer X-ray fluorescence analysis Fluorescence detection system Fluorescence microscopy Spectrofluorometer On-chip fluorescence sensor integrated an optical filter Interference filter Absorption filter Hybrid filters Filter-free fluorescence sensor Principe Light absorption characteristics Light absorption characteristics of silicon Structure and operation of the filter-free fluorescence sensor Fabrication Separation ability Performance improvement of the sensor Body-biasing technique Surface planarization of polysilicon Indium Tin Oxide photogate for detect the short wavelength Triple-well structure Application Analysis of the fluorescent reagent Quantitative cell Real-time compact identification system of amplified target RNA Conclusion Acknowledgment References

2 3 3 4 4 5 5 5 6 6 7 7 7 8 8 9 9 9 9 10 12 12 13 13 14 14 14 15 15 16 16 17 18 18

Glossary Absorption coefficient A constant indicating how much the medium absorbs when it enters a medium with light. Lab-on a chip A device that enables research that can be performed in a laboratory through a chip which has a size of a fingernail by integrating ultra-fine circuit semiconductor technology, nanotechnology, and biotechnology. Photomultiplier tube An optical sensor that provides high sensitivity by amplifying electrons emitted by the photoelectric effect. Polysilicon (polycrystalline silicon) High-purity polycrystalline silicon used as raw material in solar power generation and the electronics industry. Root mean square The arithmetic mean square root of the value of the data or the probability variable squared. Spectrum Complex information and signals are decomposed into components and are arranged according to the size of each component.

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00095-9

1

2

Physical Sensors: Fluorescence Sensors

Nomenclature S0 Ground state S1 Excited singlet state S10 Excitation singlet state f Light intensity fem Fluorescence intensity fex Intensity of excitation light a Absorption coefficient b Quantum absorption C Concentration of fluorescent reagent T Sample thickness M Molarity fr Reflected light intensity fi Incident light intensity f0 Light intensity on surface fw Light intensity at depth W W Depth R Reflectance S Area q Charge I Current h Planck’s constant c Speed of light in vacuum v Frequency l Light wavelength lem Fluorescence wavelength lex Excitation light wavelength S Sensing area gw The amount of light at depth W V (Voltage)

Abstract Fluorescence analysis is a powerful method for biochemical analyzes and medical diagnostic procedures. Various detection methods have been employed to detect fluorescence with high sensitivity. This chapter introduces the fluorescence principle and various detection methods and then explains the structures and operating principles of fluorescence microscopes and fluorescence spectrometers. A compact fluorescence sensor with high sensitivity and selectivity is required, and an on-chip fluorescence sensor incorporating a semiconductor process has been described. Finally, the structure, principle, and application of the filter-free fluorescence sensor are discussed.

Introduction Fluorescence is a type of luminescence emitted as light when a substance that absorbs ultraviolet or visible light transitions from an excited state to a ground state. Because fluorescence has a lower energy than the irradiated excited light, it is emitted at a longer wavelength than the exciting light. Fluorescence is variously observed in the natural world, such as from minerals, plants, and marine organisms, and is widely used to detect and analyze target substances in various fields because of its high applicability (Marshall and Johnsen, 2017). In particular, the fluorescence detection method has numerous advantages, such as high sensitivity and selectivity and relatively easy measurement even at low concentrations. Owing to its advantages, it is used in various fields such as chemistry, environmental sciences, medicine, biotechnology, and life sciences (Sahoo, 2012). Fluorescence spectroscopy can be used to monitor water, soil, and air quality in the environmental fields. Fluorescence detection enables the monitoring of groundwater and surface water (Gone et al., 2009; Khadka et al., 2020), petroleum hydrocarbon pollution (De Domenico et al., 1994), pesticide runoff (Nsibande and Forbes, 2016), and domestic wastewater (Galapate et al., 1998). Furthermore, X-ray fluorescence

Physical Sensors: Fluorescence Sensors

3

makes it possible to detect solid particles in atmospheric aerosols (Schmeling, 2001) and polluted soil (Peinado et al., 2010). Fluorescence is also used in the industrial field for quality control of dairy products (Shaikh and O’Donnell, 2017) and agricultural products (Zhang et al., 2012a, 2012b). Fluorescent antibiotics in the field of medicine enable the detection of antibiotics in the environment (Kümmerer, 2003), tracking antibiotic uptake in cells and throughout the organism (Kaya et al., 2003), and detecting bacterial infections (Stone et al., 2018). Fluorescence detection has become an essential analytical tool in biotechnology and life sciences with various fluorescence probes. Because of its high selectivity and sensitivity to the target, it is helpful for the analysis of biological responses (Rahmati et al., 2020), cell quantification (Alyassin et al., 2009), virus detection (Navarro et al., 2015), and nervous system of the brain (Hitti and Siegelbaum, 2014; Zhang et al., 2012a, 2012b). Fluorescence detection systems, such as fluorescence microscopy or fluorescence spectrometers, have high sensitivity and selectivity for trace amounts of target substances. However, fluorescence analysis requires expensive large-scale systems and complex expertise. Therefore, efforts such as miniaturization, cost reduction, and automation of existing devices are continuously being made to overcome these problems. Lab-on-a-chip (LOC)dwhich integrates a chemical analysis device into a miniaturized chipdaims to realize point-of-care testing (POCT) for stable and rapid analyte detection (Pol et al., 2017; Zhang et al., 2017). Because LOCs are manufactured based on microfluidics technology, a small amount of reagent can flow through a microchannel to perform complex experiments simultaneously. There are various advantages of LOCs such as a shorter reaction time, miniaturization of measuring equipment, and cost reduction. However, because large-scale devices, such as fluorescence microscopy, are widely used to analyze reactants in LOCs, the detection portion of the LOC becomes large, thus reducing the benefits associated with LOCs. Therefore, a miniaturized on-chip fluorescence detection device should be developed for measurements such as infield diagnosis using micro total analysis systems (m-TASs) and LOCs (Xu et al., 2009). Various studies have been actively conducted to develop miniaturized, low-cost, and high-resolution fluorescence detection systems for real-time detection, such as POCT. In this chapter, we present various methods for detecting fluorescence in the biofield. Basic information regarding fluorescence and its measurement methods is provided, and the structure and principle of a fluorescence microscope and fluorescence spectrometer generally used for fluorescence detection are explained. A miniaturized on-chip fluorescence sensor with an integrated optical filter for application in POCT is described. In addition, the principle and basic structure of a filter-free fluorescence sensor that can simultaneously detect a specific wavelength without an optical filter are described.

Fluorescence Principle Fluorescence is the light emitted by electrons that absorb visible and ultraviolet light and return to the ground state after excitation (Jablonski, 1933). The relationship between excitation and fluorescence is illustrated using the Jablonski diagram shown in Fig. 1A. (1) Visible and ultraviolet light is absorbed and molecules are turned to the excited states. The molecules become unstable with high energy. This state is called the excitation singlet state S10 . (2) The excess energy is lost as thermal or vibration energy and becomes the lowest energy level in the excited singlet state S1. (3) Energy is radiated as fluorescence from the excited singlet state S1 and returns to the ground state S0. The transition of the fluorescence spectrum to the long-wavelength side indicated that the energy level was lower than that of the excited state. Excess vibrational energy is converted into thermal energy when excess energy returns to the ground state. Stokes shifts are usually measured as the energy difference between absorption and maximum emission. Fig. 1B shows the relationship between excitation and fluorescence spectra, where Ex and Em are the excitation light and fluorescence, respectively. The dashed line is the excitation spectrum, indicating the absorbance of fluorescent substances. The excitation spectrum forms a smooth curve because it reflects the electronic state of the fluorescent material, and the various states of the many electrons absorb their respective energies. Fluorescence spectra are also smooth curves owing to the transition between multiple energy states. At this time, even if the excitation wavelength is changed, the peak height (light intensity) of the fluorescence spectrum changes, but the peak fluorescence wavelength does not change. The fact that the wavelength of the peak fluorescence is fixed means that even when excitation light of

Fig. 1 (A) Simple Jablonski diagram illustrating (B) the relationship between the fluorescence spectrum and excitation spectrum by Stokes’s shift and (C) the relationship between the fluorescence substance concentration and fluorescence intensity.

4

Physical Sensors: Fluorescence Sensors

various wavelengths is irradiated, the fluorescence intensity can be detected by measuring only light of a certain wavelength. In general, the Stokes shift is measured as the difference in the energy between the absorption and maximum emissions (Vollmer et al., 1994).

Fluorescence intensity Fluorescence intensity is an important parameter that can be used to measure the concentration of a target in a fluorescence detection method. The following five factors determine fluorescence intensity: (1) The fluorescence intensity (fem) is proportional to the intensity of excitation light (fex). (2) The absorption coefficient a of the fluorescent material is defined as the optical concentration of 1 M fluorescent reagent per 1 cm path. (3) The quantum absorption b of the fluorescent material is expressed as the ratio of the number of photons absorbed to the number of photons emitted and is a maximum of 1. Quantum absorption is defined as the number of photons emitted and absorbed. (4) The concentration of the fluorescent reagent in the sample is expressed as C. The fluorescence intensity and concentration are proportional in the low concentration range. (5) The sample thickness or depth of the incident excitation light is T. As shown in Fig. 1C, proportional relationships cannot be established due to concentration quenching and fluorescence reabsorption in high-concentration areas; therefore, they are usually used in proportionally low concentration ranges. Therefore, the intensity of fluorescence fem is expressed by the following formula: fem ffex $a$b$C$T

(1)

Time-resolved fluorescence lifetime In the fluorescence intensity detection method, the detection value changes according to parameters such as the concentration of the fluorophore, intensity and wavelength of the excited light, and the optical system. Therefore, it is necessary to correct the fluorescence intensity for quantitative evaluation of the analysis target. To correct these defects, a fluorescence lifetime detection method has been proposed (Lakowicz et al., 1992; Suhling et al., 2015). Time-resolved fluorescence detection is a method of measuring the fluorescence lifetime, which is an intrinsic property of a fluorophore. This approach does not depend on the fluorophore concentration, sample thickness, fluorescence intensity, or excitation intensity. The fluorescence lifetime generally ranges from picoseconds to milliseconds depending on the material and transition process (Cherry et al., 1980). The excitation stops when the intensity is reduced to 1/e of the initial value by fluorescence or non-radiative processes. Time-resolved fluorescence detection is used to calculate the fluorescence lifetime by analyzing the spectrum obtained by measuring the decay time of the emitted light by excitation of a material by a pulsed laser. Fig. 2A shows a schematic diagram of the fluorescence intensity and the time-resolved fluorescence. The fluorescence lifetime has a unique value depending on the type of material, and the excitation light pulse is shorter than the decay time of the fluorescence signal. Although the value changes depend on the measurement conditions, such as temperature or solvent, the same sample under constant conditions shows the same luminescence lifetime. Fluorescence lifetime measurement is becoming increasingly important as an analytical tool in various fields, and structural changes of proteins using fluorescence resonance energy

Fig. 2 (A) Schematic diagram of the fluorescence lifetime measurements in the time domain. Fluorescence is emitted by the excitation pulse and exponentially decreases. The fluorescence lifetime is detected by calculating the signal decay curve. (B) Schematic diagram of the fluorescence polarization analysis system. The fluorophore is excited by polarized light and depolarized by rapidly rotating small molecules bound to the fluorophore. In contrast, immobilized molecules have reduced rotation and therefore usually emit polarized light.

Physical Sensors: Fluorescence Sensors

5

transfer (FRET) (Hunt et al., 2012), pH measurement (Orte et al., 2013), and glucose detection (Saxl et al., 2009) have been reported.

Fluorescence polarization The measurement techniques using fluorescence polarization analysis methods can be used for protein-protein interactions (Rossi and Taylor, 2011), high-throughput screening of small molecules (Smith and Eremin, 2008), and drug discovery (Burke et al., 2003). Fig. 2B shows a schematic diagram of fluorescence polarization detection. The fluorescently labeled sample is irradiated with excitation light polarized through the polarization filter. The polarization direction of the fluorescence emitted by the excitation light is inversely proportional to the molecular rotation. Fluorescently labeled small molecules rotate at high speeds due to Brownian motion (Weber, 1953). Therefore, most fluorescence is emitted as depolarization. In contrast, most of the emitted fluorescence maintains the polarization state in the same direction as the excitation light because the rotational speed decreases when the fluorescently labeled detection target binds to a large molecule. Fluorescence emitted from the sample is detected in two directionsdparallel and perpendicular to the polarization plane of the excitation light. The detection intensity can be used to analyze the interactions between molecules according to the polarization direction. Therefore, since the antibody immobilization process can be omitted in the immunoassay, the measurement time is shortened, and the operation becomes simple. The fluorescence polarization method is widely used for high-throughput screening of molecules because it has the advantage of real-time analysis of association or dissociation (Zorrilla et al., 2008) and quantification analysis at low concentrations.

Fluorescence resonance energy transfer FRET was reported in 1948 by a German physical chemist called Förster (1948). It refers to the energy transfer phenomenon between two types of fluorescent molecular substances located very close to each other. This energy transfer process occurs through resonance and depends on the distance between the fluorescent molecules. The first fluorescent molecular material of the donor receives energy and becomes excited and transitions to the second fluorescent molecular material of the acceptor. Therefore, a donor for a molecule that provides energy and an acceptor for a luminescent molecular material that receives energy is required, which is located at a close distance between 1 and 10 nm (Wallace and Atzberger, 2017). As shown in Fig. 3A, FRET is satisfied only when the donor’s spectrum emits the excited energy and the spectrum that the acceptor absorbs overlap. As shown in Fig. 3B, FRET does not occur because the spectra do not overlap, and the photons emitted by the donor do not have adequate energy to be absorbed by the acceptor. In general, FRET can be detected by decreasing the fluorescence intensity of the donor or by increasing the fluorescence of the acceptor. Therefore, fluorescence can be detected by measuring the light intensity or fluorescence lifetime. FRET is being applied to visualize or understand phenomena in the field of biotechnology. For example, it has been applied to identify mechanisms for protein–protein interactions (Margineanu et al., 2016) and visualize the tertiary structure of proteins (Zhao et al., 2015). In addition, it is possible to detect cancer, evaluate the efficacy of drugs (Lu and Wang, 2010), and visually evaluate the transport of nanosized drug systems (Li et al., 2017).

X-ray fluorescence analysis X-ray fluorescence analysis (XRF) is a method for qualitative and quantitative analysis of elements using secondary X-rays generated by X-ray irradiation of a sample (Franzini et al., 1972). XRF can quickly and precisely analyze the elemental composition of a large number of samples non-destructively. In addition, XRF is an essential analysis method for environmental analysis (Kalnicky and Singhvi, 2001) and research and development of materials and products (Gallen et al., 2014) because sample preparation is simple and the analysis accuracy is high (Ida and Kawai, 2004). When the analyte is irradiated with X-rays, the inner shell electrons of the atom are excited and emitted, creating a hole in the inner shell. When the outer shell electrons move in the hole of the inner shell, the energy corresponding to the difference in the level of the electron energy of the element is emitted as an electromagnetic wave in

Fig. 3 Schematic diagram of FRET: (A) FRET occurs when the emission spectrum of the donor overlaps the absorption spectrum of the acceptor and the distance between the two molecules is within 10 nm. (B) In contrast, if the intermolecular distance is long and the emission spectrum of the donor does not overlap with the absorption spectrum of the acceptor, FRET does not occur.

6

Physical Sensors: Fluorescence Sensors

the X-ray region and is called a fluorescent X-ray. Because the fluorescent X-ray spectrum is obtained from an intrinsic element, the element constituting the sample from the spectrum can be qualitatively analyzed. In addition, since the intensity of fluorescent Xrays is proportional to the concentration of elements in the sample, quantitative analysis is possible based on the intensity.

Fluorescence detection system Fluorescence microscopy Fluorescent labeling methods are generally based on reactive derivatives of fluorophores that selectively bind to functional groups contained in target biomolecules and are widely used in biotechnology because of their non-destructive properties and the high sensitivity of fluorescence techniques (Sahoo, 2012). Fluorescence microscopy is an essential device in medical and biological fields because it can selectively observe the fluorescence wavelengths of labeled targets (Herman et al., 2001; Lichtman and Conchello, 2005). Fig. 4A shows the simplified structure and optical path of a fluorescence microscope, which consists of a light source, filter cube, objective lens, and detector. The light source used in the fluorescence microscope is generally a high-brightness light source, such as a xenon or mercury lamp (Aswani et al., 2012). These two types of arc lamps were selected based on the measurement object. The xenon lamp has the advantage that the wavelengths in the ultraviolet, visible, and infrared rays are relatively widely distributed, and the mercury lamp has the advantage of having strong excitation light intensity at some wavelengths. The recent trend in the microscope market is that LEDs are being replaced by xenon and mercury lamps because LEDs have many advantages in terms of high intensity, near-infrared emission, power consumption, and heat generation compared to conventional light sources (Herman et al., 2001). The filter cube is an optical component configured to detect the fluorescence of a specific wavelength in a fluorescence microscope and consists of an excitation filter, a dichroic mirror, and an emission filter. The excitation filter excites the sample by transmitting only light in the wavelength band. Dichroic mirrors, also called beam splitters, reflect the light passing through the excitation filter and transmit the fluorescence emitted from the sample by the excitation light. The emission filter is also called a fluorescence filter, and because the dichroic mirror does not reflect 100% excitation light, some of the excitation light is transmitted through the mirror. It blocks the excitation light once more and transmits only the fluorescence in the wavelength band that the observer wants to detect. In general, filter cubes for detecting a specific fluorescent reagent are sold as a set, and filters in a specific band can be replaced and used according to the wavelength of the excitation light and fluorescence of the fluorescent reagent. The objective lens of a microscope is essential for detecting fluorescence with high sensitivity. The objective lens of a microscope is essential for detecting fluorescence with high sensitivity, and the performance is generally determined by the values of magnification and the numerical aperture (NA) (Piston, 1998). The principle of magnifying the image of an object is to magnify the first magnified real image obtained by placing an objective lens with a short focal length close to the object with the eyepiece. The magnification of the microscope is calculated as the product of the magnification of the objective lens and the magnification of the eyepiece. The resolution of objective lenses depends on the NA. The value of NA is determined by the angle of light entering the lens and the medium between the lens and the sample. The medium between the lens and the sample is usually air, but if the NA value is increased using an emulsion of oil or water, the fluorescence can be detected with better resolution (Mouroulis and Green, 2018). Because the amount of light entering the lens increases as the diameter of the lens increases, it is preferable to use an objective lens with high resolution and a high NA value in a fluorescence microscope.

Fig. 4 Structure and optical path of (A) the fluorescence microscope and (B) spectrofluorometer. The movement path of the light source is indicated in numerical order.

Physical Sensors: Fluorescence Sensors

7

Spectrofluorometer Spectroscopy is a major analytical method used to investigate the composition of matter and related processes by studying the interaction of light with an object (Mouroulis and Green, 2018). In particular, a spectrofluorometer that can measure both absorption and emission simultaneously can be used for quantitative and qualitative analysis of samples and is used in various fields such as life sciences, medicine, and environmental sciences (DeRose et al., 2007; Manning et al., 2008; Swanson and Huffman, 2018). Because the luminous intensity of the emitted fluorescence is proportional to the intensity of the irradiated light of the sample, a quantitative value can be obtained by comparing the luminance of a reference sample of known concentration and that of a sample of unknown concentration. In addition, because the spectrofluorometer can obtain two spectra of absorption and fluorescence, qualitative analysis of samples can also be performed based on the shape of the wavelength. In general, because biomaterials have different intensities of fluorescence spectra depending on the sphere structure, the structure and reaction of biomaterials can be predicted. Fig. 4B shows the simple structure of a spectrofluorometer and consists of a light source, two monochromators, and a sample and detector. As with fluorescence microscopes, the light source uses a mercury or xenon arc lamp. The light emitted from the light source is irradiated to the grating diffraction or prism of the excitation light monochromator, and the light is scattered at different angles depending on the wavelength (Brown and Tarrant, 1978; Follath, 2001). The dispersed light is used as excitation light by adjusting the slit to select a wavelength and irradiating the sample through the integrated lens. The fluorescence emitted in all directions from the sample is usually measured at an angle of 90 degrees to minimize light interference. The sample is injected into a fluorescence cuvette made of a material such as quartz, glass, or PMMA, which is relatively light transparent. The grating of the emission monochromator disperses the emitted fluorescence, and the luminosity of the wavelength to be measured can be detected by adjusting the slit. The target wavelength is detected using a photomultiplier tube (PMT) or a photodiode, and the structure of the spectrometer can replace the emission monochromator and detector parts. The spectrofluorometer has the advantage of being able to measure the fluorescence spectrum emitted from a specific sample with high sensitivity (Spring, 1991). However, there are disadvantages in that it is challenging to apply POCT because of the large number of optical components and complex structures, and it is challenging to perform cell structure analysis because only one point of the sample is measured.

On-chip fluorescence sensor integrated an optical filter To detect or image the fluorescence intensity with a specific wavelength in a miniaturized system, studies are being conducted to integrate an optical filter that passes only a specific wavelength on a photodiode, charge-coupled device (CCD), or complementary metal-oxide semiconductor (CMOS) image sensor. Fig. 5 shows a schematic diagram of a sensor with an integrated optical filter. An optical filter is an optical component that transmits specific wavelengths of a spectrum and blocks or absorbs other wavelengths. Optical filters can be classified into interference filters and absorption filters based on the filter method. As shown in Fig. 5A, the interference filter passes specific wavelengths and reflects the others. Because the interference filter is made of one or more layers of materials with different refractive indices, it has the advantage in that the passband of a specific wavelength can be adjusted by changing the structures of the layers. In an interference filter, light passing through a low refractive index is reflected by a medium with a high refractive index. Because the transmitted wavelength changes according to the incidence angle, a proper optical system configuration is required. As shown in Fig. 5B, the absorption filter transmits only a specific wavelength by absorbing light through the filter according to the absorption characteristics of the material used. Absorption filters typically consist of colored glass or gelatin. The ability of an absorption filter to block light is based on the physical thickness of the filter and the amount of dye or pigment present. It has the advantage that the pass characteristics are maintained regardless of the angle of incident light.

Interference filter Various studies have reported that an optical interference filter is integrated on a chip to detect a specific wavelength, and only the fluorescent component is detected. According to a report by E. Thrush et al., in the fluorescent component, SiO2/Si3N4 is laminated on a CMOS image sensor (Thrush et al., 2003). The wavelengths from 450 nm to 520 nm emitted from fluorescein (lex: 494 nm,

Fig. 5

Schematic diagram of a fluorescent sensor with an integrated (A) interference filter and (B) absorption filter.

8

Physical Sensors: Fluorescence Sensors

Fig. 6 (A) Fluorescence detection sensor with a-Si:H structure with an integrated interference filter. (B) CMOS fluorescent image sensor for bio insertion with integrated absorption filter. (C) Fluorescence sensor of hybrid filter composed of an interference filter, an absorption filter, and a fiber optical plate.

lem: 521 nm) are shielded by approximately 99%, and detection up to a concentration of 85 mM is successful. According to Kamei et al., ZnS/YF3 is laminated on an a-Si:H photodiode and is combined with a microchannel and a condenser lens to realize a capillary electrophoresis apparatus (Kamei et al., 2014). It has been reported that a fluorescence sensor laminated with SiO2/Ta2O5 observes fluorescein up to 7 nM and could continuously separate 4174 bacteriophage DNA (25 ng/mL) 11 times. Fig. 6A shows a cross-sectional schematic diagram and a photomicrograph of the sensor. Because a relatively low-temperature process can manufacture the interference filter, it can be integrated with the circuit on-chip. However, the interference filter requires a laminated structure with several layers to obtain the desired characteristics, complicating the manufacturing process. The transmitted wavelength may change depending on the angle of the light incident on the sensor. Furthermore, because the detectable fluorescence wavelength is fixed, there is a problem in that it cannot cope with changes in the fluorescence reagent.

Absorption filter An absorption filter uses a substance that selectively absorbs light at a specific wavelength, cuts the absorption wavelength, and allows light of another wavelength to pass through. The absorbed light is emitted as heat, and its spectral characteristics are inferior to those of the interference filter. However, because the film-formed layer of the absorption filter is formed as a single layer, the manufacturing process is relatively easy compared to the interference filter. In addition, because the allowable value of the incident angle is more comprehensive than that of the interference filter, it is often used for devices for bio insertion (Ng et al., 2006). Yildirim et al. proposed a simple method and property analysis of an absorption filter with high wavelength band selectivity and transmittance by mixing a solubilizing dye (2RLN) and an ultraviolet curable photopolymer (NOA60) (Yıldırım et al., 2017). Sunaga et al. developed a device that combines a CMOS image sensor and an LED chip (l ¼ 470 nm) for in vivo fluorescence imaging (Sunaga et al., 2016). In Fig. 6B, the CMOS image sensor and the LED light source of the device are integrated with the absorption filter and irradiated with excitation light from the LED, and the fluorescence emitted from the living body is detected. Fluorescence was successfully detected at concentrations up to 1 nM concentration for the fluorescent reagent fluorescein-4-isothiocyanate (FITCI). Because a filter is used in the same way as an interference filter, there are some problems such as the simultaneous detection of multiple fluorescent probes and the inability to respond to changes in fluorescent reagents.

Hybrid filters A new hybrid filter has been proposed that compensates for the shortcomings of interference and absorption filters. It is designed to take advantage of the strengths of each filter technology and offset its weaknesses (Richard et al., 2009). Fig. 6C shows a fluorescence detection image sensor that uses interference and absorption filters (Sasagawa et al., 2019). By designing an adequate thickness of the filter layer, the performance can be improved by taking advantage of the interference filter, which can realize high wavelength selectivity. They also demonstrated a high-sensitivity lens less fluorescence imaging device with a wide field of view, using a hybrid

Physical Sensors: Fluorescence Sensors

9

band-pass filter composed of an interference filter, an absorption filter, and a fiber optical plate (FOP), suggesting high excitation light rejection properties. The hybrid filter was fabricated as a substrate on an FOP, coupled with a large image sensor with an imaging area of 67 mm2, to observe brain slices of green fluorescent protein transgenic mice and detect fluorescent cell bodies with a lens less imaging device.

Filter-free fluorescence sensor Fluorescence detection sensors using optical filters, such as interference filters and absorption filters, are highly sensitive to fluorescence detection, but optical filters for each fluorescent reagent are required. Multiple wavelengths cannot be detected simultaneously, and filters integrated on the chip cannot cope with changes in fluorescent reagents. Maruyama et al. proposed a filter-free fluorescence detection sensor that does not require an optical filter with a photogate structure (Maruyama et al., 2006). In principle, there is no limit to the number of wavelengths that can be detected because the multi-stage PN junction photodiode structure can be reproduced by applying a photogate voltage. High-sensitivity fluorescence detection can be achieved by adjusting the photogate voltage according to the fluorescence wavelength to be detected. Therefore, it can be used for applications that cannot be realized by the fluorescence detection principle of existing optical filters. In addition, standard CMOS processes can be used in manufacturing, and integration with peripheral circuits is easy. In this chapter, the fluorescence detection principle and application of the filter-free fluorescence sensor are explained in detail.

Principe Light absorption characteristics When light is incident on a material, a part of the light is reflected from the material’s surface, and the material absorbs light that has entered inside. The light intensity f at depth Wand light intensity fw at depth W are shown in the following equation using the reflectance R, depth W, absorption coefficient a, incident light intensity fi, and incident light intensity f0 at depth 0. These relationships are shown in Fig. 7. fr ¼ R$fi

(2)

fw ¼ f0 eaW

(3)

Moreover, because the wavelength of the fluorescent probe (420–875 nm) is within the intrinsic absorption range of the absorption characteristics of the silicon semiconductor, it can be said that the silicon semiconductor is also helpful for fluorescence detection. Filter-free fluorescence sensors focus on the differences in the absorption characteristics of light wavelengths. Based on the absorption coefficient of silicon and on Eqs. (2) and (3), the attenuation rate when silicon is irradiated can be calculated (Green and Keevers, 1995).

Light absorption characteristics of silicon Fig. 8 shows the difference in the absorption depth according to the wavelength when silicon is irradiated (Green, 2008). Shortwavelength light is absorbed near the surface, and long-wavelength light penetrates deep from the surface of silicon. Next, the number of electrons generated by the absorption of light to a certain depth W is discussed. As shown in Fig. 8, the excitation light at 470 nm attenuates 18% of the light intensity on the surface when it enters 1 mm. In other words, 82% of the excitation light is absorbed by 1 mm. Based on this idea, Eq. (2) can be rewritten as the amount of light absorbed to a certain depth W, resulting in Eq. (4). Furthermore, dividing Eq. (4) by the photon energy, Ep, gives the number of electrons per unit area generated by light absorbed up to depth W. The current I generated by light absorbed up to depth W is obtained by multiplying the prime charge q by area S to absorb light. This is shown in Eq. (6):   (4) gW ¼ f0 1  eaW

Fig. 7

Motion when light is irradiated onto the medium (A) reflection and (B) absorption of light.

10

Physical Sensors: Fluorescence Sensors

Fig. 8

Absorption depth of silicon which changes according to the light wavelength.

Ep ¼ hv ¼ h I¼ 

c l

 f0 qSl  1  eaW hc

(5)

(6)

where gw: the amount of light at depth W, f0: light intensity on the silicon surface, a: absorption coefficient, h: Planck’s constant, c: the speed of light in vacuum, v: frequency, l: light wavelength, q: electric element, and S: the sensing area. Suppose that only wavelength l of the incident light is known. In this case, if the current generated from the surface to W can be measured, the unknown parameter in Eq. (6) is only the light intensity f0, and the light intensity f0 can be calculated from the measured current value. Next, we extend this idea to consider the case in which two types of light with different wavelengths (fluorescence and excitation light) are incident. The current I1 generated from the surface to depth W1 is expressed by Eq. (7). I1 ¼

  qSl1  qSl2  1  ea1 W1 fex þ 1  ea2 W1 fem hc hc

(7)

where I1: current generated from the surface to the depth W1, fex: excitation light intensity, fem: fluorescence intensity, a1: excitation light absorption coefficient, a2: fluorescence absorption coefficient, l1: excitation light wavelength, and l2: fluorescence wavelength. Furthermore, current I2 obtained by changing the depth from the surface on which the current is measured to W2 is expressed by Eq. (8). I2 ¼

  qSl1  qSl2  1  ea1 W2 fex þ 1  ea2 W2 fem hc hc

(8)

In Eqs. (7) and (8), if the wavelength l1 $ l2, the absorption coefficient a1 $ a2, and the potential peak position W1 $ W2 are known, the only unknown parameters are the intensity of incident light fex $ fem. Therefore, it is possible to calculate the light intensities fex $ fem for each incident wavelength by solving Eqs. (7) and (8) as simultaneous equations. As a supplement, the solutions of the simultaneous Eqs. (7) and (8) are shown in Eqs. (9) and (10).     I1 1  ea2 W2  I2 1  ea2 W1 fex ¼ qSl (9) 1 a1 W1 Þð1  ea2 W2 Þ  ð1  ea1 W2 Þð1  ea2 W1 Þg hc fð1  e fem ¼ qSl hc

    I1 1  ea1 W2 þ I2 1  ea1 W1 2

fð1  ea1 W1 Þð1  ea2 W2 Þ  ð1  ea1 W2 Þð1  ea2 W1 Þg

(10)

Structure and operation of the filter-free fluorescence sensor The filter-free fluorescence sensor adopts a photogate structure, and Fig. 9 shows the basic configuration of the filter-free fluorescence sensor (Blanksby and Loinaz, 2000). In the filter-free fluorescence sensor, a p-well layer is formed on an n-type silicon substrate, and a photogate is arranged in the sensing area. The structure is such that an nþ diffusion layer is arranged adjacent to the photogate as a read terminal. At this time, the potential depth can be changed by the voltage applied to the photogate. In addition, the p-well was set to the ground level, and a positive bias was applied to the n-substrate to form a potential bend due to the PN

Physical Sensors: Fluorescence Sensors

Fig. 9

Fig. 10

11

Schematic cross-section and potential distribution of the filter-free fluorescence sensor.

Schematic diagram of the potential distribution due to sensor operation.

junction. Photoelectrons absorbed on the surface side of the potential peak depth W are collected on the surface by the mountainshaped potential and are detected as a current from the readout electrode. Fig. 10 shows the operating principle of the filter-free fluorescence sensor and the potential distribution generated by applying different voltages. First, a 0 V bias is applied to the photogate and a positive bias is applied to the n-type substrate. At this time, the substrate is irradiated by excitation light and fluorescence, and this state is the initial state. When a positive voltage V1 is applied to the photogate, the potential distribution in the sensing area becomes mountain-shaped, and the depth W1 for reading the photocurrent is determined. When the sensing area is irradiated with excitation light and fluorescence in this state, photoelectrons are generated by the photoelectric effect. The generated photoelectrons flow toward the silicon surface side, with the potential peak W1 as the boundary. Then, the photocurrent I1 can be obtained by measuring the photoelectrons flowing to the surface of silicon. Next, when the positive voltage V2 is larger than the state in which the voltage V1 is applied, the photogate potential peak position becomes position W2, which is deeper than position W1. In this state, the photocurrent I2 can be obtained by performing an exact measurement. By substituting the photocurrents I1 and I2 into Eqs. (9) and (10) and solving each equation as simultaneous equations, the incident light intensities fem and fex can be calculated.

12

Physical Sensors: Fluorescence Sensors

Fabrication The filter-free fluorescence sensor was manufactured at the solid-state functional devices research facility at Toyohashi University of Technology using a CMOS 5 mm, 1-poly 2-metal process. The fabrication process of the sensor is shown in Fig. 11A. The sensing area was 200  200 mm2, and the potential depth was adjusted in the range of 0–2 mm by the voltage applied to the photogate. An n þ diffusion layer was placed adjacent to the photogate to serve as a photoelectron outlet. In addition, the photogate and the diffusion layer were surrounded by a p-well contact to set the p-well potential and served as the outlet for holes generated in the p-well. The substrate contacts are arranged around the p-well to set the substrate potential and act as an outlet for photoelectrons created deeper than the potential saddle. By shielding light outside the sensing area with a second layer of Al, it inhibits photoelectron generation outside the sensing area. Microscopic images of the filter-free fluorescence sensor are shown in Fig. 11B.

Separation ability Fig. 12A shows the setup used to measure the separation ability of the fluorescence sensor (Choi et al., 2018). The wavelength separation ability refers to the ratio that can separate the light intensity of the excitation light and fluorescence simultaneously on the sensing area. To reduce the intensity error, the irradiated light was minimized using a microscope. Two optical fibers were connected to the microscope for irradiation purposes, and light sources with two wavelengths were used: 470 nm to mimic the excitation light and 530 nm for fluorescence. The full width at half maximums (FWHMs) of the irradiated light sources 470 and 530 nm were 25 and 32 nm, respectively. The connected light sources passed through the 5  objective lens and were irradiated with a diameter of 80 mm onto the sensing area. The standard intensity of light was measured using an optical power meter. An x-axis stage was used to move the filter-free fluorescence sensor and an optical power meter to minimize the error among measurements. The electrical characteristics of the sensor were measured using a semiconductor parameter analyzer. The light intensity at 530 nm was varied during the measurements, while the light intensity at 470 nm was maintained constant. The potential depth was changed by controlling the photogate voltage during irradiation. The separation ability was defined as the ratio between the lowest light intensity at 530 nm, which was validated by calculation at 470 nm. In the measurement, the potential depths W1 and W2 of the absorbed light were adjusted by applying voltages of 1 and 3 V to the photogate. At these voltages, the depths of the potential peaks W were 1.15

Fig. 11

(A) Microscopic image of the filter-free fluorescence sensor and (B) fabrication process.

Physical Sensors: Fluorescence Sensors

13

Fig. 12 (A) Wavelength separation ability measurement system. (B) Calculated result of the wavelength separation ability (lex: 470 nm, lem: 530 nm).

and 1.52 mm, respectively. Here, the maximum intensity of the 530 nm light and the 470 nm light incident on the sensor were set to 5000 nW/cm2, and the output current was measured while gradually decreasing the intensity of the 530 nm light from 5000 to 0.1 nW/cm2. The intensities f1 and f2 for the 470 and 530 nm light, respectively, were determined by solving the simultaneous equations obtained by substituting the measured current into Eqs. (9) and (10). Fig. 12B shows the separation ability of the filter-free fluorescence sensor using 470 and 530 nm LED light sources. The separation ability of the excitation light (470 nm) and fluorescence (530 nm) was 1.200:1.

Performance improvement of the sensor Body-biasing technique The separation ability of the filter-free fluorescence sensor was degraded owing to the defect-isolated region of the potential distribution, which is a flat and broad potential peak, as schematically represented in Fig. 13A (Moriwaki et al., 2015). The potential peak separates the photoelectrons, but in the defective separation region, the photoelectrons can cross the potential peak under the influence of thermal energy at room temperature. Therefore, the separation ability of the filter-free fluorescence sensor leads to a reduction in the sensitivity with inaccurately measured photocurrent and potential peaks. To improve the performance of filter-free fluorescence sensors, we proposed a body-biasing technique. For low photogate voltages, the defect isolation area is large because the depletion layer due to the surface potential does not expand. To compensate for this, the substrate voltage was modified to extend the depletion layer created by the p-n junction between the p-well and the n-type base. As shown in the schematic diagram of Fig. 13B, a sharp potential distribution was formed by applying a positive substrate bias voltage to improve the dynamic range and performance.

Fig. 13

Conceptual scheme of the potential distribution: (A) defective separation area and (B) variable substrate voltage.

14

Physical Sensors: Fluorescence Sensors

Surface planarization of polysilicon To improve the performance of the filter-free fluorescence sensor, polysilicon surface planarization was performed (Choi et al., 2016). A change in the optical path incident on the silicon substrate is expected depending on the surface used as the gate electrode material. As shown in Fig. 14C, the surface of the photogate was rough, and the performance was expected to decrease owing to light scattering. For surface planarization, deposition was performed by low-pressure chemical vapor deposition, and the temperature was changed from 620 to 550  C. Fig. 14 shows the polysilicon surface observed by scanning electron microscopy. The root mean square (RMS) surface roughness of the deposited polysilicon measured by atomic force microscopy was reduced from 16.29 to 1.63 nm. As shown in Fig. 14B, the proposed polysilicon surface reduces the scattering of incident light on the silicon substrate with a smaller RMS roughness. As a result, the depth of the potential peak W and the photocurrent generated between the gate electrodes increases. This indicates that the separation ability of the filter-free fluorescence sensor is increased by the planarization of polysilicon, which is the sensing area.

Indium Tin Oxide photogate for detect the short wavelength Because the filter-free fluorescence sensor uses polysilicon as the material for the photogate, detecting near-violet wavelengths is challenging. A photogate with indium tin oxide (ITO) was proposed to detect near-ultraviolet light (Teshima et al., 2019). In the transmission spectrum of ITO at a wavelength of 400 nm, the transmittance of the near-ultraviolet region was approximately 90%. Therefore, measurement of near-ultraviolet light is possible using ITO as the gate electrode material. In other words, the performance improvement of the filter-free fluorescence sensor in the near-ultraviolet region can be expected by applying the ITO electrode.

Triple-well structure Fig. 15 shows a new deep n-well design by changing from an n-type substrate to a p-type substrate in the triple-well structure. The other part of the sensor has the same structure as the conventional structure. By controlling the voltage of the deep n-well, the same characteristic improvement effect as that of the conventional structure can be expected. Furthermore, by providing the n-well in a region different from the deep n-well of the sensor, the peripheral circuit’s characteristics do not change, and the sensor and peripheral circuit can be integrated. It is also possible to measure the light passing through the saddle-point W with a photocurrent. Therefore, a higher quantum efficiency can be expected than that of n-substrate sensors. Furthermore, because the saddle-point depth can be freely changed while maintaining the potential distribution steeply by the photogate voltage and body-biasing, further high sensitivity can be expected by optimizing the light wavelength to be detected. The filter-free fluorescence sensor measures the light intensity using the absorption coefficient a of silicon according to the wavelength of light instead of removing the optical filter.

Fig. 14

Polysilicon surface observed by scanning electron microscopy. (A) RMS: 16.29 nm and (B) RMS: 1.63 nm.

Fig. 15

Sectional schematic diagram and electrostatic potential graph of the filter-free fluorescence sensor with a triple-well structure.

Physical Sensors: Fluorescence Sensors

15

The light irradiated on the silicon surface is absorbed inside silicon to generate electron–hole pairs, and the output photocurrent can be measured. Eq. (11) shows the photocurrent IPG generated based on the depth W. The light intensity f0 on the silicon surface can be calculated by substituting the measured photocurrent. Eq. (12) shows the photocurrent In  well transmitted through depth W. The light intensity f0 on the silicon surface can be calculated by substituting the measured photocurrent, as follows: IPG ¼  Inwell ¼ 

 f0 qSl  1  eaW hc

 f0 qSl  aW  eaWpn e hc

(11)

(12)

Application Analysis of the fluorescent reagent The detection performance for quantitative measurements of fluorescent dyes was analyzed using FITC and Texas Red. The filter-free fluorescence sensor can be used to calculate the light intensity for a single wavelength using Eq. (6). However, as the width of the FWHM increases, a reduction in the sensitivity and errors in the calculation are expected. To measure the intensity of the weak light generated and light with spectrum width using a fluorescent reagent, accurate calculations of the intensity are required. The filter-free fluorescence sensor consists of four layers (SiO2, polysilicon, thermal SiO2, and silicon). Because the thickness of each layer differs depending on the process, the wavelength intensity of the light source incident on the silicon differs. Consequently, accurately deriving the depth and nonlinear absorption coefficient a of the light incident on silicon can be challenging. A calculation error and a decrease in the sensitivity of the light intensity is expected. Therefore, a parametric method is required to minimize the absorption coefficient and the effect of the constants on a light source with a spectrum width. This study proposed a method for calculating the light intensity using parameters based on the output photocurrent of the sensor versus the wavelengths in the range from 450 to 625 nm; except for the current I and light intensity f components in Eqs. (7) and (8), parameters A, B, C, and D can be expressed as A¼ 

  qSl1  qSl2  1  ea1 W1 ; B ¼  1  ea2 W1 ; hc hc

C¼ 

  qSl1  qSl2  1  ea1 W2 ; D ¼  1  ea2 W2 hc hc

(11)

The components of the output current can be represented by the sum of the respective currents, as shown in Eq. (12). IW1 ¼ A•f1 þ B•f2 ; IW2 ¼ C•f1 þ D•f2 Thus, the light intensities f1 and f2 can be calculated using an inverse matrix. #1 " #   " A B IW1 f1 ¼ f2 IW2 C D

(12)

(13)

Fig. 16A shows the calculated parameters and spectra of the FITC and Texas Red solutions. The FITC parameter values can be represented by the contact points of the photogate voltages of 2 and 4 V located at the peaks of the excitation light (470 nm) and fluorescence (520 nm). Similarly, Texas Red parameter values can be obtained at the peaks of the excitation light (550 nm) and fluorescence (610 nm). Fig. 16B shows the measurement setup for assessing the device using a semiconductor parameter analyzer. To evaluate FITC and Texas Red, 4 mL of solution was deposited into disposable cuvettes. Excitation light with wavelengths of 470 and 550 nm (FITC and Texas Red) was irradiated on the side of the cuvettes. Then, a semiconductor parameter

Fig. 16 (A) Calculated parameters and spectra of the FITC and Texas Red solution. (B) Schematic view of the measurement system using FITC and Texas Red.

16

Physical Sensors: Fluorescence Sensors

analyzer was used to control the sensor and measure the photocurrent. Excitation light with wavelengths of 470 and 550 nm was then irradiated at intensities of 10 and 4 mW/cm2, respectively. At the same time, the concentrations of FITC and Texas Red in the solution were individually varied from 0.01 to 50 mM. Each fluorescent dye was prepared using diluted ethanol solution. Fig. 17 illustrates the photocurrent of the sensor when voltages of 1 to 4 V are applied to the photogate. The current values are based on five experiments. In comparison to the excitation light, the fluorescence was relatively weak, and when the concentration of FITC increased, the photocurrent decreased. When excitation light with a wavelength of 470 nm was used, 90% of the light was absorbed into the silicon substrate at a depth of 0.8 mm. Above a concentration of 50 mM, the FITC solution was saturated. The excitation light and fluorescence intensities were determined by solving the equations using the photocurrents and parameters with photogate voltages of 1 and 4 V. The calculation results for the FITC solution are shown in Fig .17B. The excitation light was reduced from 10 to 3 mW/cm2 depending on the concentration, and the fluorescence at 520 nm was detectable up to a concentration of 0.01 mM. As a result, the minimum detection level was determined to be in the sub micromolar concentration range. This result is similar to the limit of detection for the FITC solution of the fluorescence sensors, which was developed by integrating the interference filter and the absorption filter.

Quantitative cell A quantitative cell system using a filter-free fluorescence sensor has been reported using astrocytes labeled with calcein-AM (Choi et al., 2017). To measure the labeled cells, a photocurrent-based parameter was devised to minimize the effect of the absorption coefficient. Calcein-AM is an impermeable compound that exhibits strong yellow-green fluorescence (lex ¼ 490 nm, lem ¼ 515 nm). Fig. 18B shows the calculated values of the excitation light and fluorescence intensity according to the distribution of each cell. The light intensity was calculated by substituting the photocurrent detected for each cell into Eq. (13). The calculated results showed that as the number of cells increased, the fluorescence intensity increased, and the excitation light intensity decreased. Depending on the number of cells, the excitation light and fluorescence intensities from ⓐ to ⓔ showed changes of 70 and 22 mW/cm2, respectively. From this result, it is possible to measure and separate two lights simultaneously using a filterfree fluorescence sensor without using an optical filter. Fig. 18C shows a comparison of the measurement data obtained using a filter-free fluorescence sensor and a fluorescence microscope. The fluorescence microscopy data were based on the photos shown in Fig. 18A, and the brightness values of the green channel were extracted. The measured data were normalized based on the minimum and maximum cell values. The results of the filter-free fluorescence sensor were consistent with the conventional fluorescence microscopy results according to the fluorescence intensity of the labeled cells.

Real-time compact identification system of amplified target RNA A system for identifying amplified target ribonucleic acid (RNA) without an optical filter using a filter-free fluorescence sensor has been proposed (Choi et al., 2019). When RNA is amplified, the reagent is turbid by magnesium pyrophosphate, and the scattered excitation light is increased (Ishiguro et al., 2003). Therefore, interference occurs in detecting the fluorescence intensity, which is weaker than the excitation light. To solve these problems, a parametric method proposed for excitation light focuses on a threecolor LED light source and the absorption depth of the light. Fluorescent components can be detected by changing the respective photogate voltages (1 and 4 V) to remove only the excitation light component. The curve shape of the graph can determine the positive or negative RNA, and Fig. 19 shows the measurement results. In positive cases, the fluorescence increased between approximately 5 and 10 min, and a curve of a different shape was determined to be negative for the target RNA. If positive for target RNA, RNA_1 and RNA_2 are excited by blue and green LEDs, respectively. By controlling the intensity of the excitation light and the potential depth of the sensor, the excitation light and fluorescence of the reagents were successfully separated using the proposed parameters. Since the fluorescence intensity can be calculated by removing the base current component of only the excitation light, the target RNA was detected without an optical filter.

Fig. 17

(A) Output photo current, and (B) separation result of the filter-free fluorescence sensor depending on the FITC reagent concentrations.

Physical Sensors: Fluorescence Sensors

17

Fig. 18 (A) Photomicrograph of astrocytes labeled with calcein-AM. (B) Calculated intensity of excitation light and fluorescence by labeled cells with calcein-AM. (C) Comparison of the brightness intensity of the photomicrograph and the intensity of fluorescence measured by the filter-free fluorescence sensor.

Fig. 19 Calculation intensity of the fluorescence separated from the excitation light according to target RNAs. RNA_1 and RNA_2 are specifically bound to INAF probes emitting fluorescence at blue LED and Green LED, respectively. The excitation light of orange LED is used to confirm the amplification of the RNA reagent: (A) RNA_1, (B) RNA_2, and (C) RNA_Negative.

Conclusion Fluorescence detection methods are used in various applications in the field of biology. Fluorescent molecules are in high demand because of their high versatility and excellent quantitation with high sensitivity. In general, fluorescence microscopy with an integrated optical filter is mainly used to detect fluorescence. A fluorescence microscope with high sensitivity and high wavelength selectivity is used for various purposes such as cell quantification and analysis, DNA and RNA detection, protein detection, and disease detection. In addition, a fluorescence spectrometer can measure the absorption and emission spectra by scattering light. However, most of the existing equipment is expensive and large. A miniaturized fluorescence detection sensor was developed for POCT. In general, an on-chip fluorescent sensor with an absorption filter or an interference filter integrated into a CMOS or CCD image sensor has been reported. These sensors can be integrated with the circuit and can detect specific wavelengths by adjusting the optical filter. In addition, the miniaturized fluorescent sensor is expected to provide a variety of information, as it enables cellular activity in vivo. In addition, a hybrid-type fluorescent sensor with the advantages of an absorption filter and an interference filter has been reported. It has high selectivity for a specific wavelength and has the advantage of detecting fluorescence over a large area. A fluorescent sensor with a direct optical filter has high selectivity but has a disadvantage in that it is challenging to measure multiple fluorescence because the measurement wavelength is fixed. To solve this problem, a filter-free fluorescence sensor using the absorption coefficient

18

Physical Sensors: Fluorescence Sensors

of silicon has been reported. A filter-free fluorescence sensor with a photogate structure can detect photoelectrons at a specific location by controlling the potential. Accordingly, it is possible to detect two or more wavelengths simultaneously. In addition, various studies on photogate materials, surface roughness, and body-biasing technology have been conducted to increase the sensitivity to a specific wavelength. Thus, quantification of fluorescence reagents and cells and real-time detection of amplified RNA have been successful. Since the future fluorescence detection field has excellent growth potential along with semiconductor technology, it is expected that a fluorescence sensor with higher performance will be manufactured and applied in various fields.

Acknowledgment This work was supported by JST OPERA Grant Number JPMJOP1834 and JSPS KAKENHI Grant Numbers JP18H03778, JP20K14790, Japan.

References Alyassin, M.A., Moon, S., Keles, H.O., Manzur, F., Lin, R.L., Hæggstrom, E., Kuritzkes, D.R., Demirci, U., 2009. Rapid automated cell quantification on HIV microfluidic devices. Lab on a Chip 9, 3364–3369. Aswani, K., Jinadasa, T., Brown, C.M., 2012. Fluorescence microscopy light sources. Micros Today 20, 22–28. Blanksby, A.J., Loinaz, M.J., 2000. Performance analysis of a color CMOS photogate image sensor. IEEE Transactions on Electron Devices 47, 55–64. Brown, S., Tarrant, A.W., 1978. Scattered light in monochromators. Optica Acta 25, 1175–1186. Burke, T., Loniello, K., Beebe, J., Ervin, K., 2003. Development and application of fluorescence polarization assays in drug discovery. Combinatorial Chemistry & High Throughput Screening 6, 183–194. Cherry, R.J., Nigg, E.A., Beddard, G.S., 1980. Oligosaccharide motion in erythrocyte membranes investigated by picosecond fluorescence polarization and microsecond dichroism of an optical probe. Proceedings of the National Academy of Sciences 77, 5899–5903. Choi, Y.J., Takahashi, K., Matsuda, M., Hizawa, T., Moriwaki, Y., Dasai, F., Kimura, Y., Akita, I., Iwata, T., Ishida, M., Sawada, K., 2016. Filter-less fluorescence sensor with high separation ability achieved by the suppression of forward-scattered light in silicon. Japanese Journal of Applied Physics 55, 4–9. Choi, Y.J., Shinozaki, Y., Takahashi, K., Iwata, T., Koizumi, S., Sawada, K., 2017. Cell analysis system using a filter-free fluorescence sensor. In: 2017 19th International Conference on Solid-State Sensors, Actuators and Microsystems (TRANSDUCERS), IEEE, pp. 198–201. Choi, Y.J., Takahashi, K., Misawa, N., Hizawa, T., Iwata, T., Sawada, K., 2018. Multi-wavelength fluorescence detection of submicromolar concentrations using a filter-free fluorescence sensor. Sensors and Actuators B: Chemical 256, 38–47. Choi, Y.J., Takahashi, K., Noda, T., Iwata, T., Ishii, H., Sawada, K., 2019. Real-time compact identification system of amplified target RNA based on a filter-free multiple wavelengths sensor. In: 2019 20th International Conference on Solid-State Sensors, Actuators and Microsystems & Eurosensors XXXIII (Transducers & Eurosensors XXXIII), IEEE, pp. 2251–2254. De Domenico, L., Crisafi, E., Magazzù, G., Puglisi, A., La Rosa, A., 1994. Monitoring of petroleum hydrocarbon pollution in surface waters by a direct comparison of fluorescence spectroscopy and remote sensing techniques. Marine Pollution Bulletin 28, 587–591. DeRose, P.C., Early, E.A., Kramer, G.W., 2007. Qualification of a fluorescence spectrometer for measuring true fluorescence spectra. The Review of Scientific Instruments 78, 033107. Follath, R., 2001. The versatility of collimated plane grating monochromators. Nuclear Instruments and Methods in Physics Research 467–468, 418–425. Förster, T., 1948. Zwischenmolekulare Energiewanderung und Fluoreszenz. Annals of Physics 437, 55–75. Franzini, M., Leoni, L., Saitta, M., 1972. A simple method to evaluate the matrix effects in X-ray fluorescence analysis. X-Ray Spectrometry 1, 151–154. Galapate, R.P., Baes, A.U., Ito, K., Mukai, T., Shoto, E., Okada, M., 1998. Detection of domestic wastes in Kurose River using synchronous fluorescence spectroscopy. Water Research 32, 2232–2239. Gallen, C., Banks, A., Brandsma, S., Baduel, C., Thai, P., Eaglesham, G., Heffernan, A., Leonards, P., Bainton, P., Mueller, J.F., 2014. Towards development of a rapid and effective non-destructive testing strategy to identify brominated flame retardants in the plastics of consumer products. Science of the Total Environment 491–492, 255–265. Gone, D.L., Seidel, J.L., Batiot, C., Bamory, K., Ligban, R., Biemi, J., 2009. Using fluorescence spectroscopy EEM to evaluate the efficiency of organic matter removal during coagulation-flocculation of a tropical surface water (Agbo reservoir). Journal of Hazardous Materials 172, 693–699. Green, M.A., 2008. Self-consistent optical parameters of intrinsic silicon at 300K including temperature coefficients. Solar Energy Materials & Solar Cells 92, 1305–1310. Green, M.A., Keevers, M.J., 1995. Optical properties of intrinsic silicon at 300 K. Progress in Photovoltaics: Research and Applications 3, 189–192. Herman, P., Maliwal, B.P., Lin, H.J., Lakowicz, J.R., 2001. Frequency-domain fluorescence microscopy with the LED as a light source. Journal of Microscopy 203, 176–181. Hitti, F.L., Siegelbaum, S.A., 2014. The hippocampal CA2 region is essential for social memory. Nature 508, 88–92. Hunt, J., Keeble, A.H., Dale, R.E., Corbett, M.K., Beavil, R.L., Levitt, J., Swann, M.J., Suhling, K., Ameer-Beg, S., Sutton, B.J., Beavil, A.J., 2012. A fluorescent biosensor reveals conformational changes in human immunoglobulin E Fc. The Journal of Biological Chemistry 287, 17459–17470. Ida, H., Kawai, J., 2004. Identification of steel by X-ray fluorescence analysis with a pyroelectric X-ray generator. Analytical and Bioanalytical Chemistry 379, 735–738. Ishiguro, T., Saitoh, J., Horie, R., Hayashi, T., Ishizuka, T., Tsuchiya, S., Yasukawa, K., Kido, T., Nakaguchi, Y., Nishibuchi, M., Ueda, K., 2003. Intercalation activating fluorescence DNA probe and its application to homogeneous quantification of a target sequence by isothermal sequence amplification in a closed vessel. Analytical Biochemistry 314, 77–86. Jablonski, A., 1933. Efficiency of anti-stokes fluorescence in dyes. Nature 131, 839–840. Kalnicky, D.J., Singhvi, R., 2001. Field portable XRF analysis of environmental samples. Journal of Hazardous Materials 83, 93–122. Kamei, T., Sumitomo, K., Ito, S., Takigawa, R., Tsujimura, N., Kato, H., Kobayashi, T., Maeda, R., 2014. Heterogeneously integrated laser-induced fluorescence detection devices: Integration of an excitation source. Japanese Journal of Applied Physics 53. Kaya, T., Torisawa, Y.S., Oyamatsu, D., Nishizawa, M., Matsue, T., 2003. Monitoring the cellular activity of a cultured single cell by scanning electrochemical microscopy (SECM). A comparison with fluorescence viability monitoring. Biosensors & Bioelectronics 18, 1379–1383. Khadka, A., Oikonomou, C., Paudel, S.R., 2020. Real time monitoring of groundwater fluorescence: Principle and applicability in Nepal. Journal of the Institute of Engineering 15, 137–143. Kümmerer, K., 2003. Significance of antibiotics in the environment. The Journal of Antimicrobial Chemotherapy 52, 5–7. Lakowicz, J.R., Szmacinski, H., Nowaczyk, K., Berndt, K.W., Johnson, M., 1992. Fluorescence lifetime imaging. Analytical Biochemistry 202, 316–330. Li, Y., Miao, X., Chen, T., Yi, X., Wang, R., Zhao, H., Lee, S.M.Y., Wang, X., Zheng, Y., 2017. Zebrafish as a visual and dynamic model to study the transport of nanosized drug delivery systems across the biological barriers. Colloids and Surfaces. B, Biointerfaces 156, 227–235. Lichtman, J.W., Conchello, J.-A., 2005. Fluorescence microscopy. Nature Methods 2, 910–919. Lu, S., Wang, Y., 2010. Fluorescence resonance energy transfer biosensors for cancer detection and evaluation of drug efficacy. Clinical Cancer Research 16, 3822–3824.

Physical Sensors: Fluorescence Sensors

19

Manning, H.B., Kennedy, G.T., Owen, D.M., Grant, D.M., Magee, A.I., Neil, M.A.A., Itoh, Y., Dunsby, C., French, P.M.W., 2008. A compact, multidimensional spectrofluorometer exploiting supercontinuum generation. Journal of Biophotonics 1, 494–505. Margineanu, A., Chan, J.J., Kelly, D.J., Warren, S.C., Flatters, D., Kumar, S., Katan, M., Dunsby, C.W., French, P.M.W., 2016. Screening for protein-protein interactions using Förster resonance energy transfer (FRET) and fluorescence lifetime imaging microscopy (FLIM). Scientific Reports 6, 28186. Marshall, J., Johnsen, S., 2017. Fluorescence as a means of colour signal enhancement. Philosophical Transactions of the Royal Society B 372, 20160335. Maruyama, Y., Sawada, K., Takao, H., Ishida, M., 2006. A novel filterless fluorescence detection sensor for DNA analysis. IEEE Transactions on Electron Devices 53, 553–558. Moriwaki, Y., Takahashi, K., Akita, I., Ishida, M., Sawada, K., 2015. Improvement of dynamic range of filter-less fluorescence sensor with body-biasing technique. Japanese Journal of Applied Physics 54, 04DL03. Mouroulis, P., Green, R.O., 2018. Review of high fidelity imaging spectrometer design for remote sensing. Optical Engineering 57, 1. Navarro, E., Serrano-Heras, G., Castaño, M.J., Solera, J., 2015. Real-time PCR detection chemistry. Clinica Chimica Acta 439, 231–250. Ng, D.C., Tokuda, T., Yamamoto, A., Matsuo, M., Nunoshita, M., Tamura, H., Ishikawa, Y., Shiosaka, S., Ohta, J., 2006. On-chip biofluorescence imaging inside a brain tissue phantom using a CMOS image sensor for in vivo brain imaging verification. Sensors and Actuators B: Chemical 119, 262–274. Nsibande, S.A., Forbes, P.B.C., 2016. Fluorescence detection of pesticides using quantum dot materialsdA review. Analytica Chimica Acta 945, 9–22. Orte, A., Alvarez-Pez, J.M., Ruedas-Rama, M.J., 2013. Fluorescence lifetime imaging microscopy for the detection of intracellular pH with quantum dot nanosensors. ACS Nano 7, 6387–6395. Peinado, F.M., Ruano, S.M., González, M.G.B., Molina, C.E., 2010. A rapid field procedure for screening trace elements in polluted soil using portable X-ray fluorescence (PXRF). Geoderma 159, 76–82. Piston, D.W., 1998. Choosing objective lenses: The importance of numerical aperture and magnification in digital optical microscopy. The Biological Bulletin 195, 1–4. Pol, R., Céspedes, F., Gabriel, D., Baeza, M., 2017. Microfluidic lab-on-a-chip platforms for environmental monitoring. TrAC Trends in Analytical Chemistry 95, 62–68. Rahmati, M., Silva, E.A., Reseland, J.E., Heyward, A., Haugen, H.J., 2020. Biological responses to physicochemical properties of biomaterial surface. Chemical Society Reviews 49, 5178–5224. Richard, C., Renaudin, A., Aimez, V., Charette, P.G., 2009. An integrated hybrid interference and absorption filter for fluorescence detection in lab-on-a-chip devices. Lab on a Chip 9, 1371–1376. Rossi, A.M., Taylor, C.W., 2011. Analysis of protein-ligand interactions by fluorescence polarization. Nature Protocols 6, 365–387. Sahoo, H., 2012. Fluorescent labeling techniques in biomolecules: A flashback. RSC Advances 2, 7017–7029. Sasagawa, K., Ohta, Y., Kawahara, M., Haruta, M., Tokuda, T., Ohta, J., 2019. Wide field-of-view lensless fluorescence imaging device with hybrid bandpass emission filter. AIP Advances 9, 035108. Saxl, T., Khan, F., Matthews, D.R., Zhi, Z.-L., Rolinski, O., Ameer-Beg, S., Pickup, J., 2009. Fluorescence lifetime spectroscopy and imaging of nano-engineered glucose sensor microcapsules based on glucose/galactose-binding protein. Biosensors & Bioelectronics 24, 3229–3234. Schmeling, M., 2001. Total-reflection X-ray fluorescencedA tool to obtain information about different air masses and air pollution. Spectrochimica Acta Part B 56, 2127–2136. Shaikh, S., O’Donnell, C., 2017. Applications of fluorescence spectroscopy in dairy processing: A review. Current Opinion in Food Science 17, 16–24. Smith, D.S., Eremin, S.A., 2008. Fluorescence polarization immunoassays and related methods for simple, high-throughput screening of small molecules. Analytical and Bioanalytical Chemistry 391, 1499–1507. Spring, K.R., 1991. Detectors for fluorescence microscopy. Scanning Microscopy 5, 63–69. Stone, M.R.L., Butler, M.S., Phetsang, W., Cooper, M.A., Blaskovich, M.A.T., 2018. Fluorescent antibiotics: New research tools to fight antibiotic resistance. Trends in Biotechnology 36, 523–536. Suhling, K., Hirvonen, L.M., Levitt, J.A., Chung, P.-H., Tregidgo, C., Le Marois, A., Rusakov, D.A., Zheng, K., Ameer-Beg, S., Poland, S., Coelho, S., Henderson, R., Krstajic, N., 2015. Fluorescence lifetime imaging (FLIM): Basic concepts and some recent developments. Medical Photonics 27, 3–40. Sunaga, Y., Yamaura, H., Haruta, M., Yamaguchi, T., Motoyama, M., Ohta, Y., Takehara, H., Noda, T., Sasagawa, K., Tokuda, T., Yoshimura, Y., Ohta, J., 2016. Implantable imaging device for brain functional imaging system using flavoprotein fluorescence. Japanese Journal of Applied Physics 55. Swanson, B.E., Huffman, J.A., 2018. Development and characterization of an inexpensive single-particle fluorescence spectrometer for bioaerosol monitoring. Optics Express 26, 3646. Teshima, T., Sawada, K., Choi, Y.-J., Noda, Y., Kamiya, S., Akai, D., Hizawa, T., Takahashi, K., Ishii, H., Noda, T., 2019. Extension for short wavelength detection limit of filter-free fluorescence sensor by using indium tin oxide photogate. In: 2019 IEEE Sensors, IEEE, pp. 1–4. Thrush, E., Levi, O., Ha, W., Wang, K., Smith, S.J., Harris, J.S., 2003. Integrated bio-fluorescence sensor. Journal of Chromatography. A 1013, 103–110. Vollmer, F., Rettig, W., Birckner, E., 1994. Photochemical mechanisms producing large fluorescence stokes shifts. Journal of Fluorescence 4, 65–69. Wallace, B., Atzberger, P.J., 2017. Förster resonance energy transfer: Role of diffusion of fluorophore orientation and separation in observed shifts of FRET efficiency. PLoS One 12, e0177122. Weber, G., 1953. Rotational Brownian motion and polarization of the fluorescence of solutions. Advances in Protein Chemistry 8, 415–459. Xu, X., Zhang, S., Chen, H., Kong, J., 2009. Integration of electrochemistry in micro-total analysis systems for biochemical assays: Recent developments. Talanta 80, 8–18. Yıldırım, E., Arpali, Ç., Arpali, S.A., 2017. Implementation and characterization of an absorption filter for on-chip fluorescent imaging. Sensors and Actuators B: Chemical 242, 318–323. Zhang, R., Ying, Y., Rao, X., Li, J., 2012a. Quality and safety assessment of food and agricultural products by hyperspectral fluorescence imaging. Journal of the Science of Food and Agriculture 92, 2397–2408. Zhang, Z., Sharon, E., Freeman, R., Liu, X., Willner, I., 2012b. Fluorescence detection of DNA, Adenosine-50 -triphosphate (ATP), and telomerase activity by zinc(II)-Protoporphyrin IX/ G-Quadruplex labels. Analytical Chemistry 84, 4789–4797. Zhang, L., Ding, B., Chen, Q., Feng, Q., Lin, L., Sun, J., 2017. Point-of-care-testing of nucleic acids by microfluidics. Trends in Analytical Chemistry 94, 106–116. Zhao, M., Wan, X., Li, Y., Zhou, W., Peng, L., 2015. Multiplexed 3D FRET imaging in deep tissue of live embryos. Scientific Reports 5, 1–15. Zorrilla, S., Rivas, G., Acuña, A.U., Lillo, M.P., 2008. Protein self-association in crowded protein solutions: A time-resolved fluorescence polarization study. Protein Science 13, 2960–2969.

Physical Sensors: Thermal Sensors Toan Dinha,b,c, Thanh Nguyenc, Hoang-Phuong Phanc, Van Daud, Dzung Daoc,d, and Nam-Trung Nguyenc, a Centre for Future Materials, University of Southern Queensland, Springfield, QLD, Australia; b School of Mechanical and Electrical Engineering, University of Southern Queensland, Springfield, QLD, Australia; c Queensland Micro-and Nanotechnology Centre, Griffith University, Nathan, QLD, Australia; and d School of Engineering and Built Environment, Griffith University, Nathan, QLD, Australia © 2023 Elsevier Ltd. All rights reserved.

Introduction Concepts of thermal sensors Temperature sensors Thermoresistive effect Thermoresistive temperature sensors Thermal flow sensors Thermal inertial sensors Other thermal sensors Thermoelectric and pyroelectric sensors Thermoelectronic sensors Frequency analog based sensors Nanoscale effect based thermal sensors Desirable characteristics of thermal sensors High sensitivity Fast response Low power consumption High signal-to-noise ratio (SNR) Long-term stability Applications of thermal sensors Conclusion References

20 21 22 22 23 24 27 28 28 28 28 29 29 29 29 30 30 30 31 31 31

Abstract Thermal sensors have been intensively developed and used in a wide range of applications from monitoring of industrial processes to environmental control. Thermal sensors are capable of measuring various thermal signals such as temperature, flow, acceleration and angular velocity. The chapter focuses on fundamentals of physical thermal sensors that sense temperature by detecting sensor materials’ electrical properties. The chapter discusses in detail material design as well as transduction mechanisms of thermal sensors toward achieving high sensitivity. Recent advances in the development of thermal sensors are mentioned. This article provides insight into design of thermal sensors with high performance.

Introduction Detection and manipulation of thermal signals such as temperature and heat flux are of considerable interest for a wide range of applications from daily life and laboratory activities to industrial processes (Beardslee et al., 2010). Thermal sensors refer to the class of sensors utilized for measuring physical parameters based on temperature or heat flux change. Fig. 1 shows the input/output signals from thermal sensors. Thermal sensors can directly measure the temperature and heat flux changes. Thermal sensors can also be utilized to monitor indirect signals such as force, pressure, flow rate and acceleration that cause temperature changes (Kuo et al., 2012; Nguyen, 1997). Microelectromechanical system (MEMS) based thermal sensors are widely used in monitoring temperature and thermal quantities (Balakrishnan et al., 2017; Dinh et al., 2017b, 2020a). Recent development of MEMS thermal sensors has been driven toward miniaturization, high sensitivity, high frequency response and low cost (Brigante et al., 2011; Dinh et al., 2016a, 2018a). Thermal sensing materials, including metals and semiconductors, have been utilized for the development of thermal sensors (Dinh et al., 2017b). These materials have a high sensitivity to temperature and good compatibility with CMOS and MEMS fabrication technologies. For example, platinum was used to fabricate micro thermal flow sensors (Mailly et al., 2001), and lightly doped silicon was employed for convective gyroscopes (Dao et al., 2006). However, there is an emerging demand for alternative materials and designs of thermal sensors, particularly for healthcare applications. These applications require flexibility, stretchability, and wearability (Dinh et al., 2020b,c). Therefore, thermal sensors based on nanomaterials (e.g., carbon nanotube and graphene) have been recently employed to monitor temperature of body and physiological signals, providing important information for

20

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00052-2

Physical Sensors: Thermal Sensors

21

Fig. 1 Concept of thermal sensors (Nguyen, 1997). Physical signals (e.g., temperature, heat flux, flow and acceleration) are measured by thermal sensors. Various transduction mechanisms of thermal sensors are based on physical sensing effects (e.g., thermoresistive, thermoelectric, thermoelectronic, etc.).

the treatment of diseases (Dinh et al., 2016c). The requirement of thermal sensors to operate in hostile conditions (e.g., high temperature) has raised the need for high quality and low-cost materials that can operate reliably at high temperatures or high corrosion (Zhang et al., 2014). For instance, silicon carbide (SiC) thermal sensors have attracted great interest in recent years due to its excellent physical and sensing properties (Phan et al., 2019). The recent development of thermal sensors is to meet the following desirable characteristics (1) ultra-high sensitivity, (2) miniaturization, mass production and low cost, (3) high frequency, fast response and low noise, (4) excellent long-term stability in harsh environments, (5) flexibility, stretchability and wearability for emerging applications, including human health monitoring. This chapter presents the fundamentals and advanced developments of thermal sensors, including temperature sensors, thermal flow sensors and convective inertial sensors. The chapter focuses on the design of high-performance thermal sensors based on thermoresistive effect. Various desirable characteristics of thermal sensors are discussed. The chapter also discusses the recent developments of thermal sensors and provides insight into design of ultra-sensitive thermal sensors.

Concepts of thermal sensors Thermal sensors measure the change in temperature via the change of electrical properties of sensor materials. The physical input signals can be the direct temperature or indirect temperature changes such as flow and acceleration that cause the cooling effect (Sze et al., 2021). Fig. 1 shows various transduction mechanisms for thermal sensors. Several types of thermal transducers, including temperature sensors (Han et al., 2018), thermal flow sensors (Balakrishnan et al., 2018), convective accelerometers (Chaehoi et al., 2006) and gyroscopes (Dao et al., 2006), have been developed. These thermal sensors operate based on several physical sensing effects such as thermoresistance (Dinh et al., 2015c) and thermoelectricity (Milanovic et al., 1997). Thermal sensors based on thermoelectric effect have their output voltage directly converted from temperature changes. Thermoresistance or the thermoresistive effect, which refers to the change of electrical resistance corresponding to the variation in temperature, has been widely employed for thermal sensors owing to its high sensitivity and simplicity in design and fabrication. This article focuses on the fundamental and recent development of thermoresistive sensors (Dinh et al., 2017b). In thermoresistive sensors, the electrical input can be a constant current, voltage or power, while the electrical output is measured via changes of electrical resistance. When a thermal signal (e.g., temperature change) is applied to the sensor, the charge carriers in the sensor material are excited and the scattering effect is introduced, leading to the changes in the electrical resistance of the sensor (Dinh et al., 2015c). Under a constant supplied current, the change of output voltage represents the change of the input thermal signal. This sensing concept has been utilized to develop thermistors, temperature detectors and semiconductor-based integrated circuits (IC) (e.g., diodes or transistors) (Madhusoodhanan et al., 2017). In the pyroelectrical/pyroelectric and thermoelectric principles, the output signals are a voltage as a result of temperature variation and without applied current (Xu et al., 2017). Thermoelectronic sensors employ the p-n junction/diode or transistor for sensing of thermal signals (Zhu et al., 2018, 2019). In addition, frequency analog based thermal sensors utilize the shift of resonant frequency as the key sensing parameter (Guzman et al., 2020; Hsu et al., 2001). More recently, thermal sensors have been developed at nanoscale to address the need of high accuracy and spatial resolution (Yang et al., 2018; Yue and Wang, 2012).

22

Physical Sensors: Thermal Sensors

Fig. 2

Thermoresistive effects in metals and semiconductors.

Temperature sensors Temperature sensors are widely used in several practical applications thanks to their low cost, simplicity in design and implementation. Metals, semiconductors, ceramics and nanomaterials are common sensing materials in these temperature sensors. This section focuses on temperature sensors operating based on the thermoresistive effect.

Thermoresistive effect Thermoresistive effect refers to the change of electrical resistance corresponding to the change of temperature. The electrical resistance (R) of a thermoresistor is calculated from the resistivity r, the length l, width w and thickness t by R ¼ r wtl . The resistance change (dR) depends on the resistivity change (dr) due to temperature variation (dT) as follows (Dinh et al., 2017b): dR dr ¼  adT Ro ro

(1)

where a is the thermal expansion coefficient of the thermoresistor; Ro and ro are the resistance and resistivity at reference temperature To. The change of resistivity (dr ro ) is typically much higher than the change of geometric effect (adT). Therefore, the temperature coefficient of resistance (TCR) is defined based on the change of resistivity of the sensor material as: TCR ¼

dR 1 dr 1 ¼ Ro dT ro dT

(2)

Fig. 2 shows the dependence of electrical resistance on temperature in semiconductors and metals. With increasing temperature, metals exhibit an increase in electrical resistance (positive TCR), while semiconductors typically show a decrease in electrical resistance (negative TCR). The resistance of metals is described as: R ¼ Ro ½1 þ TCRðT  To Þ

(3)

The increase of resistance is attributed to the dominance of the scattering effect in metals with increasing temperature. In semiconductors, more charge carriers will be thermally excited with the rise of temperature while the mobility of charge carriers decreases. The electrical resistance of the extrinsic semiconductor material mainly depends on the activation energy (Ea) of excited charge carriers as:   Ea Rfexp (4) kT At very high temperatures, the generation of charge carriers comes from the ruptured bonds, the electrical resistance is defined as   E Rfexp 2kTg where Eg is the energy gap of the semiconductor. In highly doped semiconductors, the impurities are ionized at room temperature, the resistance can increase with increasing temperature, showing a positive TCR value which is similar to thermoresistive behaviours of metals. In semiconductor-based thermistors, the electrical resistance is expressed in the following form:    1 1 Rfexp B  (5) T To where B is the thermal index. The relationship between B and TCR can be described as TCR ¼ B T 2 (Dinh et al., 2016a). Thermoresistive sensors become popular due to the availability of thermoresistive materials, and simplicity in design, fabrication and testing of the sensors. The following section will discuss the thermoresistive effect based temperature sensors.

Physical Sensors: Thermal Sensors

23

Fig. 3 Thermoresistive effects in SiC. (A) Thermoresistive effect in different material structures including crystalline cubic SiC (3C-SiC) and amorphous structure (a-SiC) (Dinh et al., 2018a). (B) Giant TCR of up to 50%/K for SiC/Si structures under light illumination, 100 times higher than TCR measured in dark conditions. (C) Enhancement of temperature sensitivity using optoelectronic coupling effectdoptothermotronics (Dinh et al., 2020a). Dinh T, Nguyen NT and Dao DV (2018a) Thermoelectrical Effect in SiC for High-Temperature MEMS Sensors. Springer.

Thermoresistive temperature sensors Material morphologies. In thermoresistive based temperature sensors, the sensitivity depends on microstructures of the materials. For example, amorphous semiconductors have a higher temperature sensitivity (TCR) compared to extrinsic crystalline semiconductors. Fig. 3A shows a large TCR of up to  16,000 ppm/K which is almost three times higher than crystalline cubic silicon carbide (3CSiC) (Dinh et al., 2015a,c, 2018a). Different from single crystalline materials, polycrystalline ones consist of crystallites/grains and boundaries between them (Raman et al., 2006). The thermoresistive properties of crystallites are similar to that of single crystalline materials, while the grain boundaries create a significant barrier to impede the movement of electrons and govern the thermoresistive properties of the whole polycrystalline structure (Uma et al., 2001). The grain boundaries are sensitive to temperature changes due to the effect of thermionic emission and tunnelling of electrons over the boundary barriers. Selection of material morphologies (e.g., amorphous, crystalline, polycrystalline) and doping levels depend upon the requirement of sensitivity and resistivity. Amorphous and polycrystalline materials could offer a high sensitivity but also a high resistivity. Doping levels. Lowly doped semiconductors typically show a negative TCR value due to the dominance of charge carrier generation with increasing temperature (Kasap, 2006). However, highly doped semiconductors could have a small negative TCR, near zero, or positive TCR values, depending on the dominance of charge carrier generation or scattering of lattice structures (Dinh et al., 2018a). Single crystalline doped materials and metals are suitable for applications which require the thermistor/sensor with heating capability and low supply voltage.

24

Physical Sensors: Thermal Sensors

Surface and structure modifications. Several approaches have been successfully demonstrated for the enhancement of TCR for semiconductors and nanomaterials (Boragno et al., 1987; Fernandes et al., 2013; Lee et al., 2017). The modification of surface roughness of silicon by gold nanoparticles can improve the TCR value up to 100% in cryogenic temperature ranges from 10 to 30 K (Lee et al., 2017). In addition, structural modification of conductive nanotube composites through volume-phase transition can alter the tunnelling distance between the nanotubes and enhance TCR values. At the temperature of volume-phase transition, a large change in the tunnelling distance will require a high energy to excite electrons through the tunnelling barriers, resulting in a large change in electrical resistance (Fernandes et al., 2013). However, surface and structure modifications limit applications in a narrow temperature range. Physical coupling. The performance of thermal sensors depends on the natural sensing properties of materials. Recent research has successfully demonstrated the coupling of multi-physical effects in SiC/Si platform to achieve a giant TCR value which is well above the performance of natural sensing materials (Dinh et al., 2020a). The TCR value was measured to be  50%/K at non-uniform light illumination, which is one of the highest TCR value reported to date for semiconductor based temperature sensors as shown in Fig. 3B. The thermal sensitivity was increased by 100 times compared to that measured at dark conditions. Fig. 3C shows the concept of optothermotronic effect as an ultrasensitive sensing technology based on nano-heterostructure of two dissimilar semiconductors, e.g., SiC on Si. In this concept, a gradient of charge carriers is generated between two electrodes P and Q by non-uniform light illumination. Thermal excitation and transport of charge carriers are manipulated by coupling with an electric tuning current, leading to the giant temperature sensing effect that exceeds the sensing performance of the state-of-the-art semiconductor based thermal sensors.

Thermal flow sensors Thermal flow sensors are utilized to measure the velocity of surrounding fluids and the direction of flow. The working principle of thermal flow sensors is based on the heat transfer between the sensor heater and the environment, and the thermoresistive effect of the sensor material. Thermal flow sensors which have a heating element (i.e., heater) are known as hot-wire/hot-film anemometers. Anemometers require a heater with a reasonable resistivity which can raise the temperature of the heater with low supply voltage/power. Joule heating effect will drive the heater into a steady state where the temperature is higher than that of the surrounding environment. When the flow rate increases, the heater is cooled down, leading to change in the sensor’s electrical resistance. By measuring the resistance changes, the amplitude of flow rate is quantified. Fig. 4A shows the circuitry and working principle of hot-wire and hot-film flow sensors. In a constant current mode of operation, the resistance of the sensor is measured by the output voltage Vout. In the calorimetric principle, a supply voltage creates a rise in temperature of the heater in Fig. 4B. Two resistors (i.e., upstream and downstream) arranged in form of a Wheatstone bridge will detect the change in temperature profile of the heater under the change of flow rates. The temperature of the downstream resistor is typically higher than that of the upstream resistor. In the time of flight concept, the flow is measured by the transition time of heat pulse from heater to the temperature sensor as shown in Fig. 4C. The transition time depends on flow rate, thermal conductivity and diffusivity of the surrounding environment. Among three sensing concepts, hot-wire/hot-film flow sensors show advantages in term of simplicity in design, characterization and measurement of air flow. Fig. 5A shows the change of temperature profile and heat dissipation of the heater which is detected by measuring the differential voltage of the heater. Fig. 5B shows the temperature profile of a graphite hot-film flow sensor which was constructed on a paper substrate. Under 10 mA supply current, the change of output voltage was from 100 to 600 mV, depending on

Fig. 4 Measurement circuits of thermal flow sensors (Dinh et al., 2017b; Meng et al., 2008). (A) Hot-wire/hot-film principle. (B) Calorimetric principle. (C) Time of flight concept. (A): Dinh T, Phan HP, Qamar A, Woodfield P, Nguyen NT and Dao DV (2017b) Thermoresistive effect for advanced thermal sensors: Fundamentals, design considerations, and applications. Journal of Microelectromechanical Systems 26(5): 966–986; (B) Meng E, Li PY and Tai YC (2008) A biocompatible Parylene thermal flow sensing array. Sensors and Actuators A: Physical 144(1): 18–28.

Physical Sensors: Thermal Sensors

25

Fig. 5 Characterization of hot-film air flow sensors. (A) Flow rate induces the heat dissipation on the hot-film (Dinh et al., 2017a). (B) Optothermal image of the heater. (C) Measured output voltage versus time. (D) Dependence of differential voltage on the air flow velocity (Dinh et al., 2015b). (A): Dinh T, Phan HP, Nguyen TK, Qamar A, Woodfield P, Zhu Y, Nguyen NT and Dao DV (2017a) Solvent-free fabrication of biodegradable hot-film flow sensor for noninvasive respiratory monitoring. Journal of Physics D: Applied Physics 50(21): 215401; (C and D) Dinh T, Phan HP, Dao DV, Woodfield P, Qamar A and Nguyen NT (2015b) Graphite on paper as material for sensitive thermoresistive sensors. Journal of Materials Chemistry C 3(34): 8776–8779.

the flow velocity as shown in Fig. 5C. Empirical estimation of relationship between the differential voltage velocity v can be described using King’s law:

OV ¼ a þ bvn

O V and the flow (6)

where a, b and n are empirical constants. These constants are determined by fitting Eq. (6) with the experimental data as shown in Fig. 5D. The sensitivity of hot film thermal flow sensor is calculated by s ¼ b/P, where P is the power consumption of the sensor. Thermal flow sensor can measure flow rate in microfluidic applications. Fig. 6A shows a microfluidic cooling device with an inlet and an outlet. The flow rate of water is measured by the thermal flow sensor which has a heater and two temperature sensors. The flow convectively cools down the heater which in turn decreases the temperature of the temperature sensors. Fig. 6B shows the structure of the flow sensor and microfluidic system with the model of thermal conductance. Fig. 6C shows the temperature changes of the sensors, representing the processes of heating (e.g., Heating ON where temperature increases from 25  C to 113  C) and cooling (e.g., Cooling ON where the water flow is applied). Fig. 6D shows the dependence of the sensor temperature on the flow rate of the microfluidic device. The maximum temperature change is 25  C. By measuring the temperature of the device, the flow rate is monitored. Thermal flow sensors play an important role in design, measurement and control of microfluidic systems. There is an emerging demand for flow sensors to measure flow rates at high temperatures, including gas flow in a rig where the temperature can reach over 350  C (Balakrishnan et al., 2019; Dinh et al., 2019). Recently, SiC has been attracted interest for high temperature flow sensors owing to its capability of working reliably at high temperatures. Fig. 7A shows the dependence of resistance change of SiC nanofilms on supply power density. The resistance change is above 90%, showing the possibility of raising the heater temperature of well above environment temperatures. Fig. 7B shows the response of the SiC thermal flow sensor to an exhausted air with a temperature of approximately 130  C. The decrease in the measured current indicates the increase of the electrical resistance or the cooling of the SiC nanofilms under applied air flow. Semiconductor-based thermal flow sensors are conventionally utilized for a wide range of applications in industrial processes, including detection of leaks, pipe bursts or blockages and variation of liquid concentration. Recently, nanomaterial-based flow sensors have been employed in healthcare and wearable applications, owing to their flexibility, wearability and multifunctionality (Dinh et al., 2016c, 2020b,c). For example, a wearable thermal flow sensor was developed with a carbon nanotube yarn as a heater and paper as a flexible substrate, which is affixed on human philtrum to measure flow rate from human breaths (Dinh et al., 2016c). Fig. 8 shows the response of the hot-wire wearable flow sensor to human exhalation and inhalation. Based

26

Physical Sensors: Thermal Sensors

Fig. 6 Thermal flow sensor in microfluidic application (Dinh et al., 2018c). (A) 3D sketch of a microfluidic device with thermal flow sensor. (B) Structure of the device. (C) Response of the thermal flow sensor to heating and cooling process. (D) The change in temperature of the thermal flow sensor corresponding to the flow rates. Dinh T, Phan HP, Kashaninejad N, Nguyen TK, Dao DV and Nguyen NT (2018c) An on-chip SiC MEMS device with integrated heating, sensing, and microfluidic cooling systems. Advanced Materials Interfaces 5(20): 1800764.

Fig. 7 High temperature flow sensors (Dinh et al., 2018b). (A) Resistance change versus power density applied on SiC thin film. (B) Response of flow sensor to air flow at 130  C. Dinh T, Phan HP, Nguyen TK, Balakrishnan V, Cheng HH, Hold L, Lacopi A, Nguyen NT and Dao DV (2018b) Unintentionally doped epitaxial 3C-SiC (111) nanothin film as material for highly sensitive thermal sensors at high temperatures. IEEE Electron Device Letters 39(4): 580–583.

Physical Sensors: Thermal Sensors

27

Fig. 8 Application of thermal flow sensors in healthcare (Dinh et al., 2016c). Response of thermal flow sensor to human inhalation and exhalation. Dinh T, Phan HP, Nguyen TK, Qamar A, Foisal ARM, Viet TN, Tran CD, Zhu Y, Nguyen NT and Dao DV (2016c) Environment-friendly carbon nanotube based flexible electronics for noninvasive and wearable healthcare. Journal of Materials Chemistry C 4(42): 10061–10068.

on the response of the flow sensor, the breath rate and depth can be monitored, providing important information to evaluate the physical status of patients.

Thermal inertial sensors Conventional mechanical inertial sensors, including accelerometers and gyroscopes, employ the movement of proof mass to induce stress/strain on piezoresistive element and measure the acceleration. Thermal inertial sensors, known as convective accelerometers and gyroscopes, measure acceleration using thermoresistive effect, which do not require a proof mass (Luo et al., 2001). Convective inertial sensors have high shock resistance and simplicity in design and fabrication (Van et al., 2019). Fig. 9A show the working principle of convective accelerometers which utilize a heater to generate symmetry profile of temperature distribution. The inertial force generated by acceleration push hot fluid bubbles in the direction of the applied acceleration, inducing an asymmetry of temperature profile. Two identical temperature sensors (e.g., having the same material, electrical resistance and distance to the heater) measure the temperature difference on the temperature profile. The development of convective accelerometers has been driven toward miniaturization, high sensitivity, high frequency response and wide operation range. These sensors require packaging of enclosed chamber to avoid noise and disturbance from the surrounding environment. However, decrease of cavity dimension of the sensor in the enclosed chamber will decrease the sensitivity and increase the frequency response. Using gases with high thermal diffusivity and low viscosity will enhance sensitivity and frequency response. In addition, convective accelerometers with capability of detecting two dimensional (2D) (Dao and Sugiyama, 2007) and three dimensional (3D) (Dinh et al., 2018d) signals have been developed.

Fig. 9 Thermal inertial sensors (Dinh et al., 2017b). (A) Working principle of convective accelerometers (Mailly et al., 2003). (B) Working principle of convective gyroscopes (Dao et al., 2006). Dinh T, Phan HP, Qamar A, Woodfield P, Nguyen NT and Dao DV (2017b) Thermoresistive effect for advanced thermal sensors: Fundamentals, design considerations, and applications. Journal of Microelectromechanical Systems 26(5): 966–986.

28

Physical Sensors: Thermal Sensors

Fig. 10 (A) p-n thermoelectric sensor used as radiation sensor (Vieira et al., 2019). (B) Frequency analog based thermal sensor using bridge resonant structure (Guzman et al., 2020). (A): Vieira EM, Figueira J, Pires AL, Grilo J, Silva MF, Pereira AM and Goncalves LM (2019) Enhanced thermoelectric properties of Sb2Te3 and Bi2Te3 films for flexible thermal sensors. Journal of Alloys and Compounds 774: 1102–1116; (B) Guzman P, Dinh T, Phan HP, Joy AP, Qamar A, Bahreyni B, Zhu Y, Rais-Zadeh M, Li H, Nguyen NT (2020) Highly-doped SiC resonator with ultra-large tuning frequency range by Joule heating effect. Materials & Design 194: 108922.

Fig. 9B shows the working principle of convective gyroscopes, which utilize the hot air flow from a nozzle orifice. The Coriolis acceleration deflects the flow direction, causing an opposite cooling effect between the two temperature sensors. The change of resistance is typically converted to an output voltage using a Wheatstone bridge. Selection of material for thermoresistive temperature sensors plays an important role in design of highly sensitive convective sensors. For example, gyroscopes based on lightly doped silicon is tens of times more sensitive than those made by metals (e.g., tungsten).

Other thermal sensors Thermoelectric and pyroelectric sensors Pyroelectric sensors deploy the generation of temporary voltage in pyroelectric materials (e.g., gallium nitride and lithium tantalate) due to heating/cooling effect. Thermoelectric sensors employ a heating element and temperature sensing element. The sensing element includes several thermocouples of thermoelectric materials connected in series or a thermopile. Thermocouples operate based on the Seebeck effect which is the electric potential formed by a temperature gradient. The increase of thermocouples in series will increase the output voltage of the thermopile but also increase the noise. The output voltage from the thermopiles depends on the difference in Seebeck coefficients of the thermoelectric materials and thermal isolation between the hot and cold junctions (Kuo et al., 2012). The output voltage can be described as Vout ¼ aDT, where a is the Seebeck coefficient of thermocouple or the difference of Seebeck coefficient of two thermoelectric materials; DT is the temperature difference between hot and cold junction. Conventional materials for thermoelectric sensors are metals (e.g., chromium and nickel) and semiconductors (e.g., silicon). The fabrication of many thermoelectric sensors requires CMOS (complementary metal oxide semiconductor) processes which are more complicated than thermoresistive sensors. Fig. 10A shows the structure of p-Sb2Te3 and n-Bi2Te3 film based thermal sensors used to detect heat flow which warms up the thermopile (hot junction) while the temperature of the cold junctions is kept unchanged using a heat sink. The temperature difference between the hot and cold junctions generates a voltage which can be measured by a voltmeter.

Thermoelectronic sensors Thermoelectronic sensors refer to diodes and transistors as a temperature sensing element. The change of barrier of p-n junction is  qV  described as I ¼ Io enkT 1 , where q is the electron charge, k is the Boltzmann constant, and n is the ideality factor (Matthus et al., 2017; Rao et al., 2015b). To date, the development of thermoelectronic sensors has been driven toward high sensitivity in a wide temperature range (Dao et al., 2007; Rao et al., 2015b). For example, thermoelectronic sensors based on 4H-SiC p-n junctions have been demonstrated up to 600  C with excellent linearity and high sensitivity of 3.5 mV/oC (Rao et al., 2015a; Zhang et al., 2014).

Frequency analog based sensors Frequency analog based sensors, known as resonant based thermal sensors, utilize the changes of resonant frequency of mechanical structures and surface acoustic wave structures caused by the change in the temperature and subsequent thermal stress. The basic mechanical structures, including cantilevers and bridges, have been employed to detect thermal signals. The change of resonant

Physical Sensors: Thermal Sensors frequency Of f depends on the change of temperature

O T by Of f

¼ TCF 

29

OT, where TCF is the temperature coefficient of

frequency. In most mechanical structures, increase in temperature causes the decrease in the resonant frequency due to the soften mechanism of materials at elevated temperatures (Dinh et al., 2016b). The frequency analog mechanism is commonly utilized to develop reliable temperature sensors (Hsu et al., 2001). In a more recent work, a high tuning range of Of f ¼ 80% has been successfully demonstrated for SiC bridge structures (Guzman et al., 2020) as shown in Fig. 10B. This shows a potential of developing highly sensitive resonant temperature sensors.

Nanoscale effect based thermal sensors The recent development of thermal devices used in emerging fields such as biological and clinical applications has raised the need for thermal sensors to be developed at a nanoscale level (Bai and Gu, 2016). Different thermal sensing approaches at nanoscale have been developed with high precision which exceeds the conventional thermal sensors in term of accuracy and spatial resolution. At nanoscale, the heat transfer is unusual with the contribution of phonon confinement, interface scattering phenomenon (Brites et al., 2012; Cahill et al., 2003). Several nanoscale thermal sensors and techniques have been utilized including nanolithography, nanomaterials, fluorescence and superstructures (Lee and Kotov, 2007).

Desirable characteristics of thermal sensors Several desirable features are considered in designing, fabricating and implementing of thermal sensors. The desirable characteristics include high sensitivity, linear response, long-term stability, capability of operating in extreme environments including high temperature and under mechanical disturbances, and flexibility/wearability for wearable applications.

High sensitivity The sensitivity of thermal sensors characterizes the amplitude of measured signals in corresponding to an applied signal. A high sensitivity presents the accuracy and efficiency of monitoring temperature changes. Temperature coefficient of resistance (TCR), defined as the relative resistance change to the temperature variation, is the representative factor for the sensitivity of thermal sensors. An alternative indicator for sensitivity is the thermal index (B) which is commonly used to evaluate the thermal sensitivity of semiconductors and ceramic materials. A thermal sensor with a high TCR material or high B-value material usually shows a high sensitivity. For temperature sensors, a giant thermoresistive effect with a TCR value of above 5%/K could be desirable to achieve a high sensitivity of temperature detection. Ceramics and composite materials typically provide a high sensitivity to temperature with a TCR of up to 1012 ppm/K (Ohe and Naito, 1971). However, a giant TCR has been a great challenge for semiconductor thermal sensors due to the limitation of charge generation in semiconductors. Recent research has demonstrated innovative methods to enhance the performance of semiconductor based thermal sensors with TCR of up to 50%/K (Dinh et al., 2020a). Thermal flow sensors and convective inertial sensors operate based on the Joule heating effect, the resistivity is a critical factor in addition to the sensitivity. Metals and highly doped semiconductors have been widely used for these sensors thanks to their high conductivity. However, the sensitivity is limited at a TCR value of below 7000 ppm/K (Dinh et al., 2017b). A low TCR or near zero TCR is beneficial to mechanical sensors and others in terms of design and implementation of temperature control/compensation.

Fast response Response time of thermal sensors plays a key role in real-time measurement of signals. A fast response time enables thermal sensors to respond instantaneously to the change of temperature, flow and acceleration. Experimentally, the response time is determined from 63.2% (or 90% in some applications) of the output signal (Sosna et al., 2011). For example, the thermopile output signals from the thermal flow sensor (Ke et al., 2019) is estimated to be approximately 250 ms based on 63.2% output signal as shown in Fig. 11. The response time can also be estimated from exponential fit of the sensor temperature T and the time t using T ¼ A  B  e t/s, where s is the time constant; A and B are experimental constants . A typical response time of MEMS thermal sensors is in millisecond range or a response frequency in kHz ranges. Theoretically, the response of thermal sensors depends upon the thermal resistance Rth and thermal capacitance Cth of the sensing element, with the time constant determined by s ¼ L  rVC , where L and A are the length and cross-sectional area of the thermal sensing element; V and C are Rth  Cth ¼ Ak H H the volume and heat capacity of the sensing element; k and r are the thermal conductivity and density of the sensor material. To achieve a faster response, designers can consider to scale down the dimensions of the sensor or/and select materials with high thermal conductivity, low density and low heat capability.

30

Physical Sensors: Thermal Sensors

Fig. 11 Response time characteristics. Estimation of response time based on 63.2% (or 90%) of sensor output (Ke et al., 2019). Ke W, Liu M, Li T and Wang Y (2019) MEMS thermal gas flow sensor with self-test function. Journal of Micromechanics and Microengineering 29(12): 125009.

Low power consumption Design of thermal sensors with low power consumption is of interest for practical applications. Scaling down the sensor structure not only enhances the response time, but also enables a lower power consumption. The mature of micro/nanomachining technologies has led to the development of various nanostructures such as nanotube, nanowires and nanobelts which have small volume and dimensions and subsequent low power consumption (Prades et al., 2008). For temperature sensors, the power consumption is relatively low as these sensors do not require a heating element. In contrast, heater-based sensors require well-designed isolation to minimize power losses due to conduction. Sub-mW power consumption of MEMS thermal sensors based on design of suspended structures was successfully demonstrated (Cubukcu et al., 2010). A common solution is to design the heating element on a suspended membrane. The design of heating elements on a thick thermal-conducted substrate should be avoided (Dinh et al., 2019). It is important to note that an increase in the supply power will also increase the sensitivity of thermal sensors and the power losses. Convective accelerometers and gyroscopes find a wide range of applications in stabling of cell phone and camera, displays, anti-roll fields and other electronic devices.

High signal-to-noise ratio (SNR) SNR refers to the ratio between the power of the desired output signal and the background noise, which is described as SNR ðdBÞ ¼   V , where Vsignal and Vnoise are the measured signal voltage and noise voltage, respectively. Thermal sensors with a high 2log10 Vsignal noise SNR are desired for a wide range of application. Reduction of noise is used to increase the SNR. There are different sources of noise, including external (electromagnetic), conducted and intrinsic noise (Dao et al., 2004). While external and conducted noises can be removed and reduced by circuit designs and appropriate setup, intrinsic noise is unavoidable and can be reduced by appropriate design of heater and thermal sensing elements. Intrinsic sources of noise include Johnson noise (random motion of carrier) and 1/f noise (Dinh et al., 2017a). The intrinsic noise can be reduced by increasing number of carriers in semiconductors (reducing resistance) and increasing the volume of thermal sensing element.

Long-term stability Thermal sensors operating reliably in long-term services are preferable for a wide range of practical applications. The desirable characteristics are the ability of working in the environments without significant degradation of material properties or sensing performance over time. The consistency in response of sensors after thousands of testing cycles is required for thermal sensors working at room temperature or at elevated temperatures. However, the deterioration in services of sensor materials and sensing characteristics is typically observed due to the effects of absorption and desorption of water molecules. Using lowly porous materials would reduce

Physical Sensors: Thermal Sensors

31

the effect of humidity on sensor performance. In addition, to achieve the stability of sensor operation at high temperature and in harsh environments, designers can select wide bandgap materials such as silicon carbide (SiC) and diamond as these materials have been proven to show excellent stability and repeatability of working in the hostile conditions (Yang et al., 2019).

Applications of thermal sensors Thermal sensors cover a wide range of applications in monitoring and control of industrial processes, laboratory conditions, medical procedures and daily life activities. Rapid progress has been made in the development of thermal sensors in the fields of automotive, aerospace and defence. Sensing devices for direct measurements of temperature have been developed and commercialized, including resistive temperature detectors (RTDs) and thermocouples. These sensors become popular in the current temperature sensing markets thanks to their low cost, high sensitivity and wide temperature range of operation. For some industrial applications with temperature of below 600  C, RTD sensors can replace thermocouples (Feteira, 2009). RTD sensors show linear responses to temperature changes, but much lower sensitivity compared to semiconductor and ceramic based thermal sensors that have negative temperature coefficient of resistance (NTC) (Dinh et al., 2017b; Feteira, 2009). Thermal flow sensors are also popular in various industrial applications such as natural gas flow measurement, steel production, combustion control, nitrogen/air monitoring, gas valve testing and pressure regulation, etc. In these applications, thermal sensors are directly located in the gas flow for measurements of mass flow rates, densities, volumes and directions of air/gases or liquids, as well as determination of flow patterns (Ejeian et al., 2019). Convective accelerometers and gyroscopes find a wide range of applications in monitoring acceleration of vehicle, machines and buildings, and in stabling of cell phone and camera, displays, anti-roll fields and other electronic devices (Mukherjee et al., 2017; Yazdi et al., 1998). Thermal sensors become increasingly popular in biomedical and healthcare applications thanks to the biocompatibility and capability of fabricated at nanoscale levels (Ramgir et al., 2010). For example, thermal sensors have been employed for thermal characterization of skin to evaluate the subtle variations in skin temperature associated with mental health (Webb et al., 2013). The use of thermal sensors for monitoring human breaths (see Fig. 8), providing important information for detection of abnormal breath patterns and assessment of human health (Dinh et al., 2016c).

Conclusion A wide range of well-established concepts and new discoveries have driven the development of thermal sensors toward ultra-high sensitivity, miniaturization and suitability for emerging and niche applications. Many materials (metals, semiconductors and nanomaterials), sensing transduction types (thermoresistive, thermoelectric and thermoelectronic), and designs/configurations have been used to develop thermal sensors based on micro/nanomachining technologies. Thermoresistive sensors have been of great interest owing to their simplicity in design, fabrication and measurement. Thermal sensors with high shock resistance, high sensitivity, fast response and superior properties (flexible, stretchable and capable of working in hostile conditions) will continue expending their engineering applications in different fields.

References Bai, T., Gu, N., 2016. Micro/nanoscale thermometry for cellular thermal sensing. Small 12 (34), 4590–4610. Balakrishnan, V., Phan, H.-P., Dinh, T., Dao, D.V., Nguyen, N.-T., 2017. Thermal flow sensors for harsh environments. Sensors 17 (9), 2061. Balakrishnan, V., Dinh, T., Phan, H.-P., Dao, D.V., Nguyen, N.-T., 2018. Highly sensitive 3C-SiC on glass based thermal flow sensor realized using MEMS technology. Sensors and Actuators A: Physical 279, 293–305. Balakrishnan, V., Dinh, T., Nguyen, T., Phan, H.-P., Nguyen, T.-K., Dao, D.V., Nguyen, N.-T., 2019. A hot-film air flow sensor for elevated temperatures. Review of Scientific Instruments 90 (1), 015007. Beardslee, L.A., Addous, A.M., Heinrich, S., Josse, F., Dufour, I., Brand, O., 2010. Thermal excitation and piezoresistive detection of cantilever in-plane resonance modes for sensing applications. Journal of Microelectromechanical Systems 19 (4), 1015–1017. Boragno, C., Valbusa, U., Pignatel, G., 1987. Fast P-doped silicon bolometer for detecting heat pulses. Applied Physics Letters 50 (10), 583–584. Brigante, C.M., Abbate, N., Basile, A., Faulisi, A.C., Sessa, S., 2011. Towards miniaturization of a MEMS-based wearable motion capture system. IEEE Transactions on Industrial Electronics 58 (8), 3234–3241. Brites, C.D., Lima, P.P., Silva, N.J., Millán, A., Amaral, V.S., Palacio, F., Carlos, L.D., 2012. Thermometry at the nanoscale. Nanoscale 4 (16), 4799–4829. Cahill, D.G., Ford, W.K., Goodson, K.E., Mahan, G.D., Majumdar, A., Maris, H.J., Merlin, R., Phillpot, S.R., 2003. Nanoscale thermal transport. Journal of Applied Physics 93 (2), 793–818. Chaehoi, A., Mailly, F., Latorre, L., Nouet, P., 2006. Experimental and finite-element study of convective accelerometer on CMOS. Sensors and Actuators A: Physical 132 (1), 78–84. Cubukcu, A.S., Zernickel, E., Buerklin, U., Urban, G.A., 2010. A 2D thermal flow sensor with sub-mW power consumption. Sensors and Actuators A: Physical 163 (2), 449–456. Dao, D.V., Sugiyama, S., 2007. A 2-DOF convective micro accelerometer with a low thermal stress sensing element. Smart Materials and Structures 16 (6), 2308. Dao, D.V., Toriyama, T., Sugiyama, S., 2004. Noise and frequency analyses of a miniaturized 3-DOF accelerometer utilizing silicon nanowire piezoresistors. In: SENSORS, 2004 IEEE, pp. 1464–1467. Dao, D.V., Shiozawa, T., Kumagai, H., Sugiyama, S., 2006. Development of a dual-axis thermal convective gas gyroscope. Journal of Micromechanics and Microengineering 16 (7), 1301.

32

Physical Sensors: Thermal Sensors

Dao, D.V., Shiozawa, T., Sugiyama, S., 2007. Convective gas gyroscope based on thermo-resistive effect in Si pn junction. In: TRANSDUCERS 2007-2007 International Solid-State Sensors, Actuators and Microsystems Conference, IEEE, pp. 2525–2528. Dinh, T., Dao, D.V., Phan, H.-P., Wang, L., Qamar, A., Nguyen, N.-T., Tanner, P., Rybachuk, M., 2015a. Charge transport and activation energy of amorphous silicon carbide thin film on quartz at elevated temperature. Applied Physics Express 8 (6), 061303. Dinh, T., Phan, H.-P., Dao, D.V., Woodfield, P., Qamar, A., Nguyen, N.-T., 2015b. Graphite on paper as material for sensitive thermoresistive sensors. Journal of Materials Chemistry C 3 (34), 8776–8779. Dinh, T., Phan, H.-P., Kozeki, T., Qamar, A., Namazu, T., Nguyen, N.-T., Dao, D.V., 2015c. Thermoresistive properties of p-type 3C–SiC nanoscale thin films for high-temperature MEMS thermal-based sensors. RSC Advances 5 (128), 106083–106086. Dinh, T., Phan, H.-P., Kozeki, T., Qamar, A., Fujii, T., Namazu, T., Nguyen, N.-T., Dao, D.V., 2016a. High thermosensitivity of silicon nanowires induced by amorphization. Materials Letters 177, 80–84. Dinh, T., Phan, H.-P., Kozeki, T., Qamar, A., Namazu, T., Zhu, Y., Nguyen, N.-T., Dao, D.V., 2016b. Design and fabrication of electrothermal SiC nanoresonators for high-resolution nanoparticle sensing. In: 2016 IEEE 16th International Conference on Nanotechnology (IEEE-NANO), IEEE, pp. 160–163. Dinh, T., Phan, H.-P., Nguyen, T.-K., Qamar, A., Foisal, A.R.M., Viet, T.N., Tran, C.-D., Zhu, Y., Nguyen, N.-T., Dao, D.V., 2016c. Environment-friendly carbon nanotube based flexible electronics for noninvasive and wearable healthcare. Journal of Materials Chemistry C 4 (42), 10061–10068. Dinh, T., Phan, H.-P., Nguyen, T.-K., Qamar, A., Woodfield, P., Zhu, Y., Nguyen, N.-T., Dao, D.V., 2017a. Solvent-free fabrication of biodegradable hot-film flow sensor for noninvasive respiratory monitoring. Journal of Physics D: Applied Physics 50 (21), 215401. Dinh, T., Phan, H.-P., Qamar, A., Woodfield, P., Nguyen, N.-T., Dao, D.V., 2017b. Thermoresistive effect for advanced thermal sensors: Fundamentals, design considerations, and applications. Journal of Microelectromechanical Systems 26 (5), 966–986. Dinh, T., Nguyen, N.-T., Dao, D.V., 2018a. Thermoelectrical Effect in SiC for High-Temperature MEMS Sensors. Springer. Dinh, T., Phan, H.-P., Nguyen, T.-K., Balakrishnan, V., Cheng, H.-H., Hold, L., Lacopi, A., Nguyen, N.-T., Dao, D.V., 2018b. Unintentionally doped epitaxial 3C-SiC (111) nanothin film as material for highly sensitive thermal sensors at high temperatures. IEEE Electron Device Letters 39 (4), 580–583. Dinh, T., Phan, H.P., Kashaninejad, N., Nguyen, T.K., Dao, D.V., Nguyen, N.T., 2018c. An on-chip SiC MEMS device with integrated heating, sensing, and microfluidic cooling systems. Advanced Materials Interfaces 5 (20), 1800764. Dinh, T.X., Dang, L.B., Tran, C.-D., Bui, T.T., Hoa, P.T., 2018d. Tri-axis convective accelerometer with closed-loop heat source. Sensors and Actuators A: Physical 275, 51–59. Dinh, T., Nguyen, T.K., Phan, H.P., Nguyen, Q., Han, J., Dimitrijev, S., Nguyen, N.T., Dao, D.V., 2019. Thermoresistance of p-type 4H–SiC integrated MEMS devices for hightemperature sensing. Advanced Engineering Materials 21 (3), 1801049. Dinh, T., Nguyen, T., Foisal, A., Phan, H.-P., Nguyen, T.-K., Nguyen, N.-T., Dao, D., 2020a. Optothermotronic effect as an ultrasensitive thermal sensing technology for solid-state electronics. Science Advances 6 (22), eaay2671. Dinh, T., Nguyen, T., Phan, H.-P., Nguyen, N.-T., Dao, D.V., Bell, J., 2020b. Stretchable respiration sensors: Advanced designs and multifunctional platforms for wearable physiological monitoring. Biosensors and Bioelectronics 2020, 112460. Dinh, T., Nguyen, T., Phan, H.P., Nguyen, T.K., Dau, V.T., Nguyen, N.T., Dao, D.V., 2020c. Advances in rational design and materials of high-performance stretchable electromechanical sensors. Small 2020, 1905707. Ejeian, F., Azadi, S., Razmjou, A., Orooji, Y., Kottapalli, A., Warkiani, M.E., Asadnia, M., 2019. Design and applications of MEMS flow sensors: A review. Sensors and Actuators A: Physical 295, 483–502. Fernandes, G.E., Kim, J.H., Sood, A.K., Xu, J., 2013. Giant temperature coefficient of resistance in carbon nanotube/phase-change polymer Nanocomposites. Advanced Functional Materials 23 (37), 4678–4683. Feteira, A., 2009. Negative temperature coefficient resistance (NTCR) ceramic thermistors: An industrial perspective. Journal of the American Ceramic Society 92 (5), 967–983. Guzman, P., Dinh, T., Phan, H.-P., Joy, A.P., Qamar, A., Bahreyni, B., Zhu, Y., Rais-Zadeh, M., Li, H., Nguyen, N.-T., 2020. Highly-doped SiC resonator with ultra-large tuning frequency range by Joule heating effect. Materials & Design 194, 108922. Han, S., Kim, J., Won, S.M., Ma, Y., Kang, D., Xie, Z., Lee, K.-T., Chung, H.U., Banks, A., Min, S., 2018. Battery-free, wireless sensors for full-body pressure and temperature mapping. Science Translational Medicine 10 (435), eaan4950. Hsu, W.-T., Clark, J.R., Nguyen, C.T.-C., 2001. A resonant temperature sensor based on electrical spring softening. In: Transducers’ 01 Eurosensors XV. Springer, pp. 1456–1459. Kasap, S.O., 2006. Principles of Electronic Materials and Devices. McGraw-Hill New York. Ke, W., Liu, M., Li, T., Wang, Y., 2019. MEMS thermal gas flow sensor with self-test function. Journal of Micromechanics and Microengineering 29 (12), 125009. Kuo, J.T., Yu, L., Meng, E., 2012. Micromachined thermal flow sensorsdA review. Micromachines 3 (3), 550–573. Lee, J., Kotov, N.A., 2007. Thermometer design at the nanoscale. Nano Today 2 (1), 48–51. Lee, S.-H., Hwang, S., Jang, J.-W., 2017. Giant temperature coefficient of resistivity and cryogenic sensitivity in silicon with galvanically displaced gold nanoparticles in freeze-out region. ACS Nano 11 (2), 1572–1580. Luo, X., Yang, Y., Zheng, F., Li, Z., Guo, Z., 2001. An optimized micromachined convective accelerometer with no proof mass. Journal of Micromechanics and Microengineering 11 (5), 504. Madhusoodhanan, S., Sandoval, S., Zhao, Y., Ware, M., Chen, Z., 2017. A highly linear temperature sensor using GaN-on-SiC heterojunction diode for high power applications. IEEE Electron Device Letters 38 (8), 1105–1108. Mailly, F., Giani, A., Bonnot, R., Temple-Boyer, P., Pascal-Delannoy, F., Foucaran, A., Boyer, A., 2001. Anemometer with hot platinum thin film. Sensors and Actuators A: Physical 94 (1–2), 32–38. Mailly, F., Giani, A., Martinez, A., Bonnot, R., Temple-Boyer, P., Boyer, A., 2003. Micromachined thermal accelerometer. Sensors and Actuators A: Physical 103 (3), 359–363. Matthus, C.D., Erlbacher, T., Hess, A., Bauer, A.J., Frey, L., 2017. Advanced 4H-SiC pin diode as highly sensitive high-temperature sensor up to 460 C. IEEE Transactions on Electron Devices 64 (8), 3399–3404. Meng, E., Li, P.-Y., Tai, Y.-C., 2008. A biocompatible Parylene thermal flow sensing array. Sensors and Actuators A: Physical 144 (1), 18–28. Milanovic, V., Gaitan, M., Bowen, E.D., Tea, N.H., Zaghloul, M.E., 1997. Thermoelectric power sensor for microwave applications by commercial CMOS fabrication. IEEE Electron Device Letters 18 (9), 450–452. Mukherjee, R., Basu, J., Mandal, P., Guha, P.K., 2017. A review of micromachined thermal accelerometers. Journal of Micromechanics and Microengineering 27 (12), 123002. Nguyen, N., 1997. Micromachined flow sensorsdA review. Flow Measurement and Instrumentation 8 (1), 7–16. Ohe, K., Naito, Y., 1971. A new resistor having an anomalously large positive temperature coefficient. Japanese Journal of Applied Physics 10 (1), 99. Phan, H.-P., Nguyen, T.-K., Dinh, T., Qamar, A., Iacopi, A., Lu, J., Dao, D.V., Rais-Zadeh, M., Nguyen, N.-T., 2019. Wireless battery-free SiC sensors operating in harsh environments using resonant inductive coupling. IEEE Electron Device Letters 40 (4), 609–612. Prades, J., Jimenez-Diaz, R., Hernandez-Ramirez, F., Barth, S., Cirera, A., Romano-Rodriguez, A., Mathur, S., Morante, J., 2008. Ultralow power consumption gas sensors based on self-heated individual nanowires. Applied Physics Letters 93 (12), 123110. Raman, M.S., Kifle, T., Bhattacharya, E., Bhat, K., 2006. Physical model for the resistivity and temperature coefficient of resistivity in heavily doped polysilicon. IEEE Transactions on Electron Devices 53 (8), 1885–1892. Ramgir, N.S., Yang, Y., Zacharias, M., 2010. Nanowire-based sensors. Small 6 (16), 1705–1722. Rao, S., Pangallo, G., Della Corte, F.G., 2015a. 4H-SiC pin diode as highly linear temperature sensor. IEEE Transactions on Electron Devices 63 (1), 414–418. Rao, S., Pangallo, G., Pezzimenti, F., Della Corte, F.G., 2015b. High-performance temperature sensor based on 4H-SiC Schottky diodes. IEEE Electron Device Letters 36 (7), 720–722.

Physical Sensors: Thermal Sensors

33

Sosna, C., Walter, T., Lang, W., 2011. Response time of thermal flow sensors with air as fluid. Sensors and Actuators A: Physical 172 (1), 15–20. Sze, S.M., Li, Y., Ng, K.K., 2021. Physics of Semiconductor Devices. John Wiley & Sons. Uma, S., McConnell, A., Asheghi, M., Kurabayashi, K., Goodson, K., 2001. Temperature-dependent thermal conductivity of undoped polycrystalline silicon layers. International Journal of Thermophysics 22 (2), 605–616. Van, N.T., Bui, T.T., Dinh, T.X., Tran, C.-D., Phan-Thanh, H., Duc, T.C., 2019. A circulatory ionic wind for inertial sensing application. IEEE Electron Device Letters 40 (7), 1182–1185. Vieira, E.M., Figueira, J., Pires, A.L., Grilo, J., Silva, M.F., Pereira, A.M., Goncalves, L.M., 2019. Enhanced thermoelectric properties of Sb2Te3 and Bi2Te3 films for flexible thermal sensors. Journal of Alloys and Compounds 774, 1102–1116. Webb, R.C., Bonifas, A.P., Behnaz, A., Zhang, Y., Yu, K.J., Cheng, H., Shi, M., Bian, Z., Liu, Z., Kim, Y.-S., 2013. Ultrathin conformal devices for precise and continuous thermal characterization of human skin. Nature Materials 12 (10), 938–944. Xu, D., Wang, Y., Xiong, B., Li, T., 2017. MEMS-based thermoelectric infrared sensors: A review. Frontiers of Mechanical Engineering 12 (4), 557–566. Yang, F., Yang, N., Huo, X., Xu, S., 2018. Thermal sensing in fluid at the micro-nano-scales. Biomicrofluidics 12 (4), 041501. Yang, T., Chen, S., Li, X., Xu, X., Gao, F., Wang, L., Chen, J., Yang, W., Hou, X., Fang, X., 2019. High-performance SiC Nanobelt Photodetectors with long-term stability against 300 C up to 180 days. Advanced Functional Materials 29 (11), 1806250. Yazdi, N., Ayazi, F., Najafi, K., 1998. Micromachined inertial sensors. Proceedings of the IEEE 86 (8), 1640–1659. Yue, Y., Wang, X., 2012. Nanoscale thermal probing. Nanotechnology Reviews 3 (1), 11586. Zhang, N., Lin, C.-M., Senesky, D.G., Pisano, A.P., 2014. Temperature sensor based on 4H-silicon carbide pn diode operational from 20 C to 600 C. Applied Physics Letters 104 (7), 073504. Zhu, C., Chortos, A., Wang, Y., Pfattner, R., Lei, T., Hinckley, A.C., Pochorovski, I., Yan, X., To, J.W.-F, Oh, J.Y., 2018. Stretchable temperature-sensing circuits with strain suppression based on carbon nanotube transistors. Nature Electronics 1 (3), 183–190. Zhu, C., Wu, H.-C., Nyikayaramba, G., Bao, Z., Murmann, B., 2019. Intrinsically stretchable temperature sensor based on organic thin-film transistors. IEEE Electron Device Letters 40 (10), 1630–1633.

Physical Sensors: Optical Sensors Hiromasa Shimizu, Department of Electrical and Electronic Engineering and Department of Applied Physics and Chemical Engineering, Tokyo University of Agriculture and Technology, Koganei, Japan © 2023 Elsevier Ltd. All rights reserved.

Introduction Classification of optical sensors Optical sensors based on free-space optics Optical sensors based on waveguide optics Relationship between concentration of analyte and complex refractive index Dielectric-based waveguide RI sensors based on Si slot waveguides Waveguide-based RI sensors with analyte discrimination Magneto-optical SPR sensor with Pd thin film for detecting H2 gas Magneto-optical SPR sensor for discrimination of analytes based on the complex refractive index VOC sensor with SPR imaging with cross-reactive sensor microarray Summary References Relevant websites

34 35 35 36 38 39 39 41 41 44 45 47 48

Nomenclature BPF Band pass filter CCD Charge coupled device EDFA Er doped optical fiber amplifier LD Laser diode LED Light emitting diode MO Magneto optical SM coupler Single mode coupler SMF Single mode fiber VOC Volatile organic compound

Abstract The principles, classification, and recent advances in optical sensors for application to biosensors and gas sensors are described. Optical sensors based on free-space optics, and waveguide optics are introduced, and difference of the principle, operation wavelength, and size are described. Waveguide-based refractive index sensors with analyte discrimination In particular, three novel approaches in which analytes are detected with high sensitivity and discrimination without antigen– antibody reactions are introduced.

Introduction Light is used in various situations such as displays, illumination, communication, and imaging. Various optical devices have been developed and are widely available at low cost, such as light sources, photodetectors, cameras, optical waveguides, optical fibers, high-speed and highly efficient optical modulators, optical filters, and integrated optics. Coherent and incoherent light sources and optical detectors have been developed for ultraviolet, visible, near-infrared, and infrared light. Cameras with an array of photodetectors, especially CMOS cameras, enable high-resolution detection and image analysis. Recent advances in microfabrication technology and synthesis of dissimilar materials have led to the development of highly efficient and compact light sources and detectors, and optical modulators driven by small external forces. There are many kinds of interactions between light and matter, such as refraction, absorption, fluorescence, and scattering. Changes in the concentration and types of matter in the surrounding environment can be regarded as “external forces” that affect these interactions, allowing these optical instruments to be applied to physical sensors, including optical sensors. Advances in the development of optical devices have contributed to the realization of transducers in highly sensitive sensors that convert slight changes in the concentration and type of matter into large electric signals.

34

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00099-6

Physical Sensors: Optical Sensors

35

One of the advantages of optical sensors is that integrated optics allows a large number of sensors including light sources and photodetectors to be realized on a single chip. When an analyte consisting of multiple molecules is adsorbed on several receptors, antibodies, and polymers, that are modified on an optical sensor, corresponding output signals can be obtained and quantitatively analyzed. By measuring the temporal transition of the output signals, it is possible to analyze the strength and kinetics of the adsorption and desorption of the analytes. In this chapter, the principles, classification, and recent advances in optical sensors for application to biosensors and gas sensors are described.

Classification of optical sensors Optical sensors are classified into two types according to the scheme of light propagation (Yariv and Yeh, 2006). The first type includes sensors based on “free-space optics,” which involves measuring the change in transmitted light intensity in the space where the analyte is introduced. Second type includes sensors based on “waveguide optics,” which involves measuring the change in propagation constant due to differences in the concentration and type of analytes adsorbed on the sensor surface. Optical sensors based on “free-space optics” have been mainly used to measure the concentration and type of gas molecules by selecting the wavelength used for detecting target molecules. The transmitted light intensities in a reference space (gas cell) with a fixed concentration of the target molecule and in a space containing analytes are measured and compared, allowing the concentration of the target molecule to be measured. On the other hand, in optical sensors based on “waveguide optics,” analytes are chemically or physically adsorbed on part of an optical waveguide or nanostructure composed of stacked media with different refractive indices, decorated with an adsorbing layer having a thickness of 1–10 nm, such as antibodies, receptors or polymers. The change in propagation constant due to changes in the concentration and type of analyte is measured from the transmission characteristics of the optical cavity or the reflection characteristics due to differences in the coupling conditions to the optical waveguide. The state of adsorption of the analyte can be observed as a temporal change in the transmitted light intensity and the reflected light intensity, and the concentration and type are obtained by comparing the signal changes with those obtained from various analytes used as references. An optical sensor based on “waveguide optics” is smaller and more mechanically stable than one based on “free-space optics.” Enhancement of the signal by using an optical nonlinear effect can be realized in the detection of analytes with localized surface plasmon resonance with enhanced Raman scattering (SERS) (Ye et al., 2010, p. 163106). A list of various optical sensors based on free space optics and waveguide optics are shown in Table 1.

Optical sensors based on free-space optics Functional groups (C-O, O-H, C-H groups, etc.) in organic compounds constantly vibrate, and the bond angle is changing (angular vibration). These functional groups vibrate at unique frequencies and show different frequencies depending on the vibration Table 1

A list of various optical sensors based on free space optics and waveguide optics. Optical sensors based on free space optics

Optical sensors

Analyte

References

NDIR

CO2, CO, CH4, NO, and SO2

CRDS

CO2, H2O, H2S, NH3

FLRDS

CO2

Hodgkinson et al. (2013), 580, https://www.yokogawa. com/us/solutions/products-platforms/processanalyzers/gas-analyzers/ftnir-ir/ Tarsa et al. (2004), 297, https://www.picarro.com/ company/technology/crds Tong et al. (2004), 6594 and Shimizu and Noriyasu (2014), 116601

Optical sensors based on waveguide optics Optical sensors Metal-based plasmonic waveguide sensor SPR sensor with SERS Dielectric-based waveguide sensor

Subcategory 1. 2.

Propagating surface plasmon polariton Localized surface plasmon polariton

1 2

Dielectric wire waveguide Dielectric slot waveguide

References Xu et al. (2019), 1801433 Ye et al. (2010), 163106 Barrios (2009), 4751 Robinson et al. (2008), 4296 Tomono and Shimizu (2019), JTh2A.96

36

Physical Sensors: Optical Sensors

method. Light having a frequency corresponding to the unique frequency of the functional group is absorbed. Absorption occurs when the dipole moment of a functional group changes periodically due to vibration. Therefore, the periodic change of the dipole moment of the molecule due to the molecular vibration results in the absorption of light. Usually, the absorption of light by molecular vibration occurs in the infrared region at wavelengths longer than 2 mm, which are shown in the HITRAN database (http://www.cfa.harvard.edu/HITRAN/). The wavelengths of light absorbed by molecular vibrations are summarized in the HITRAN database. It is possible to measure the gas concentration of CO2, CO, NO, NO2, N2O, SO2, NH3, CH4 etc. in the concentration range of several ppm to several % by using non-dispersive infrared (NDIR) optical sensors (Hodgkinson et al., 2013, 580, https://www.yokogawa.com/us/solutions/products-platforms/process-analyzers/gas-analyzers/ftnir-ir/). Generally, as the measurement wavelength becomes shorter, the optical absorption becomes smaller, so that the required optical path (size of the sensor) becomes longer. It is possible to increase the equivalent optical path length by introducing gas molecules into a resonator or cavity composed of a pair of mirrors. Optical sensors based on cavity ring down spectroscopy (CRDS) are commercially available, and detection of NH3 at 0.1 ppb has been reported (Tarsa et al., 2004, p. 297, https://www.picarro.com/company/technology/crds). Also, it is possible to increase the equivalent optical path length by introducing the gas molecules into a resonator composed of an optical fiber loop, a technique known as fiber loop ring-down spectroscopy (FLRDS) (Tong et al., 2004, p. 6594; Shimizu and Noriyasu, 2014, p. 116601). Fig. 1 shows a schematic diagram of the measurement principle of FLRDS for detecting CO2 concentration at a wavelength of 1572.334 nm (Shimizu and Noriyasu, 2014, p. 116601). CO2 concentrations of 0–0.5% were detected with a measurement resolution of 0.05%. Fig. 2 shows the ring-down waveforms for CO2 concentrations of (a) 0%, (b) 0.05%, and (c) 0.5%. Fig. 3 shows the CO2 concentration dependence of the signal (ring down time) for calculated and experimental results. The light from an infrared laser can be transmitted by an optical fiber, which is useful for remote continuous measurement of gas concentration. In optical sensors based on free-space optics, it is important to avoid the influence of optical absorption by water vapor contained in the atmosphere.

Optical sensors based on waveguide optics Polarization of the electric field of light by matter is related to the refraction and absorption of light and is expressed by the complex refractive index. Analytes such as proteins, DNA, glucose, and gas molecules have their own complex refractive indexes, depending on the wavelength of light. Generally, as the molecular weight of the analyte becomes larger, the complex refractive index becomes larger (National Astronomical Observatory of Japan, 2019). The refractive index (RI) of a mixture composed of more than two kinds of molecules varies depending on the composition ratio. Therefore, optical sensors for biosensors or gas sensors can be realized by measuring the difference in complex refractive index and converting it to an electrical signal. The change in RI upon the change in concentration increases as the molecular weight increases, so that high RI sensitivity is required for detection of analytes having smaller molecular weight, such as gas molecules, compared with the detection of macromolecules, such as proteins. Light is refracted or reflected by dielectrics, semiconductors, and metals. Optical sensors for biosensors or gas sensors are realized by confining the light into the region of the adsorbed analyte and measuring the transmitted light intensity or reflected light intensity. Light propagates through the medium having a higher refractive index; however, the RI of the analytes, solvent, and surrounding atmosphere is smaller than of that of typical dielectrics and semiconductors (Yariv and Yeh, 2006). Therefore, the important point for realizing higher sensitivity is to confine the light into the region of adsorbed analytes having smaller refractive index. One method is to introduce the analyte into the low-index layer surrounding the core in optical fibers (optical waveguides) where the light is confined and propagated by means of the difference in refractive index (Xu et al., 2019, p. 1801433). By measuring the difference in propagation constant, the concentrations and types of analytes can be measured. Light waves can propagate in

Output pulse

Input pulse Iin(t)

With absorption material

absorptive material

Iout(t) SMF SMF

l [m]

1×2 SM Coupler

SMF

EDFA

t=0

t=τ

1×2 SM Coupler

BPF

SMF

t=0

t=τ0

Without absorption material

Fig. 1 Schematic diagram of the measurement principle of fiber-loop ring-down spectroscopy. Adapted from Shimizu H and Noriyasu H (2014) Measurement of carbon dioxide concentration by fiber-loop ring-down spectroscopy for continuous remote measurement. Japanese Journal of Applied Physics 53: 116601, with permission Copyright (2014) The Japan Society of Applied Physics.

Physical Sensors: Optical Sensors

37

(A)

(B)

(C)

Fig. 2 Ring-down waveforms for CO2 concentrations of (A) 0%, (B) 0.05%, and (C) 0.5%. Adapted from Shimizu H and Noriyasu H (2014) Measurement of carbon dioxide concentration by fiber-loop ring-down spectroscopy for continuous remote measurement. Japanese Journal of Applied Physics 53: 116601, Copyright (2014) The Japan Society of Applied Physics.

38

Physical Sensors: Optical Sensors

Fig. 3 CO2 concentration dependence of the calculated and experimentally obtained D(1/s) summarized from Figs. 2A–4C. Adapted from Shimizu H and Noriyasu H (2014) Measurement of carbon dioxide concentration by fiber-loop ring-down spectroscopy for continuous remote measurement. Japanese Journal of Applied Physics 53: 116601, Copyright (2014) The Japan Society of Applied Physics.

association with collective oscillations of electrons at the interface between a metal and dielectric, and the light is confined in a narrow region smaller than the wavelength, a phenomenon known as a surface plasmon polariton (SPP) (Xu et al., 2019, p. 1801433). The difference in refractive index can be measured by introducing the analyte as a dielectric medium on the metal surface. At the interface between the metal nanoparticles and surrounding dielectric, localized surface plasmons are generated on the metal surface in a region with a size of tens of nm, and exhibit a unique absorption wavelength owing to the polarization of free electrons inside the metal nanoparticle. The absorption wavelength depends on the refractive index of the medium near the metal nanoparticles. The change in refractive index can be measured by the difference in transmitted light intensity and the spectrum. The propagation constant (equivalent refractive index) increases with increasing concentration of the analyte (the analyte brakes the propagating light). So far, various refractive index sensors have been reported, such as metal-based plasmonic waveguide RI sensors and dielectricbased waveguide RI sensors. Comparisons have been reported from the viewpoint of the wavelength of light, the sensitivity (RI resolution), and figure of merit (FOM) determining the measurement accuracy (Xu et al., 2019, p. 1801433). The sensitivity is determined by the optical confinement factor of the region of the adsorbed analyte on the optical waveguide sensor. The FOM increases as the propagation loss of the optical waveguide decreases. Therefore, the FOM of dielectric-based waveguide RI sensors is larger than that of metal-based plasmonic waveguide RI sensors, which show a larger propagation loss due to the metal. A longer wavelength is more useful for achieving lower propagation loss and higher FOM in metal-based plasmonic waveguide RI sensors. On the other hand, since the propagating light spreads over the entire dielectric layer, not only the analytes adsorbed on the waveguide surface but also the change in the refractive index of the entire dielectric layer (analytes that do not bind to the surface) is measured, leading to a decrease of the signal to noise ratio. The change of RI due to the introduction of the analyte onto the metal surface (the size of the region in which surface plasmons are localized is typically hundreds of nm) can cause a signal larger than that caused by specific binding of the analyte. This is a longstanding problem especially in metal-based plasmonic waveguide RI sensors with high RI resolution, and the problem has been overcome by differential surface plasmon resonance sensor using a quadrant cell photodetector (Zhang et al., 2003, p. 150). A longer wavelength is useful for achieving better RI resolution in dielectric-based waveguide RI sensors because the propagating light spreads in the low-refractive-index layer as an evanescent wave. The length of the light distribution region depends on the refractive index profile of the waveguide; it is several mm in SiO2-based waveguides including optical fiber, and tens of nm in Si/ SiO2-based waveguides (Xu et al., 2019, p. 1801433).

Relationship between concentration of analyte and complex refractive index Here, changes in the complex refractive index caused by differences in the concentrations and types of analytes such as aqueous glucose solution, DNA, various other solutions, and gas molecules are described in order to discuss the refractive index resolution required for biosensors and gas sensors. The refractive indexes of an aqueous glucose solution at a wavelength of 486.13 nm are reported to be 1.34768 and 1.35497 for glucose concentrations of 0 and 100 g L 1 (Yeh, 2008, p. 666). The refractive index of DNA at a wavelength of 632.8 nm was reported to range from 1.51 to 1.56 (Samoc et al., 2007, p. 236). The refractive indexes of ethanol and water at 20  C are reported to be 1.3618 and 1.3330, respectively, at a wavelength of 589.3 nm (National Astronomical Observatory of Japan, 2019, p. 441). The refractive index of ethanol gas molecules in the atmosphere is 1.1  10 4 RI$(vol%) 1 (Shimizu et al., 2021, p. 724528). The absorption coefficient of water is 0.70 m 1 (extinction coefficient of 3.9  10 8) at a wavelength of 700 nm (Pope, and Fry, 1997, p. 8710). The optical absorption by electronic transitions of molecules is more enhanced in ultraviolet to far ultraviolet light (wavelength of 120–400 nm) compared with visible light (wavelength of 400–700 nm) (Nobre et al., 2008, p. 550; Ito et al., 1969, p. 2453).

Physical Sensors: Optical Sensors

39

Dielectric-based waveguide RI sensors based on Si slot waveguides Dielectric wire waveguides provide strong optical confinement into a narrow core (< 1000  1000 nm) owing to high index contrast in a Si/Si3N4/SiO2 (or air) system, leading to more compact integrated optics. Si slot waveguides have been reported as unique optical elements that confine the propagating light into a narrow slot section (width wslot: 50–200 nm) having a lower index, between parallel Si wires. Therefore, Si slot waveguides have been applied to biosensors (Barrios, 2009, p. 4751) and gas sensors (Robinson et al., 2008, p. 4296; Tomono and Shimizu, 2019, JTh2A.96). Since a Si/Si3N4/SiO2 system is chemically stable, Si slot waveguides are suitable for on-chip sensing. In order to improve the sensitivity, enhancing the FOM (Q factor) in a ring resonator is effective for accurate readout of the resonance wavelength to detect the refractive index (RI) difference Dn. At the same time, it is possible to detect Dn through measuring the transmission intensity. It is reported that an RI difference Dn of 1.5  10 4 can be detected in a Si3N4/SiO2 system (Barrios, 2009, p. 4751). It is also reported that the sensitivity was enhanced by 3.5 times by coating biotin receptor molecules to detect avidin molecules in solutions (Claes, et al., 2009, p. 1943) in a Si/SiO2 system. On the other hand, there have been a limited number of reports on gas sensing with Si slot waveguides. C2H2 has been detected by a Si slot waveguide with the slot width (wslot) of 50 nm (Robinson et al., 2008, p. 4296). CO2 has been detected by a Si slot waveguide with wslot ¼ 200 nm (Tomono, and Shimizu, 2019, JTh2A.96). Narrowing the slot (wslot ¼ 50 nm) makes it possible to detect C2H2 gas and realize a sensitivity of 490 nm$RIU 1 (RI difference Dn ¼ 3.2  10 4) by enhancing the optical confinement factor in the slot section (G ¼ 0.64) (Robinson et al., 2008, p. 4296). Confining light into a gas molecule adsorption layer (thickness: 1– 10 nm) coated on the surface of the slot is important for realizing the detection of the adsorbed gas molecules without undesired signals by the analyte introduced in the middle of the slot section and degradation of the signal to noise ratio. Fig. 4 shows the optical mode profile (distribution of the propagating light) of the Si slot waveguide with (a) wslot ¼ 50 nm and (b) wslot ¼ 200 nm, calculated by the finite-difference-frequency-domain method (Kaihara, and Shimizu, 2017, p. 730). The width and thickness of the Si layer are set to 300 and 220 nm, respectively. The propagating light is confined at the slot section and the intensity is almost constant for wslot ¼ 50 nm (Fig. 4A), with the optical confinement factor in the slot G ¼ 0.71. On the other hand, the propagating light is confined at the slot section, and the intensity is stronger at the surface of the slot than that in the middle of the slot for wslot ¼ 200 nm (Fig. 4B), with G ¼ 0.37. In the Si slot waveguide with wslot ¼ 200 nm, the light is confined at the slot surface and is not confined at the middle of the slot. Overlapping of the gas adsorption layer at the slot surface and the distribution of the propagating light is larger for wslot ¼ 200 nm than that for wslot ¼ 50 nm. Therefore, a Si slot waveguide with slot width wslot ¼ 200 nm is more suitable for detection of the binding the molecules with the adsorption layer, compared with those having a narrower slot width (50 nm), where the light intensity is almost constant in the slot section, and intensity is not negligible in the middle of the slot. Fig. 5 shows an optical microscope image of the Si slot waveguide ring resonator with a ring radius of 200 mm. Fig. 6 shows cross-sectional scanning electron microscope images of the fabricated Si slot waveguide. Fig. 7A shows the optical transmission spectra for the device in air measured with a wavelength-tunable laser and an optical power meter. The transmission spectra with CO2 and dry-air atmospheres are measured as shown in Fig. 7B. The resonant wavelength shifts to the longer wavelength side by 0.04 nm under CO2. Since the RIs of CO2 and dry air are 1.00029 and 1.00044, respectively, Dn is 1.5  10 4, corresponding to a sensitivity of 3  102 nm$RIU 1.

Waveguide-based RI sensors with analyte discrimination In biosensing, a change in the concentration of the target molecule can be measured by a sensor chip decorated with antibodies or DNA that specifically binds to a target molecule. On the other hand, specific binding of gas molecules is difficult. In this section,

Fig. 4 Calculated optical mode profiles of a Si slot waveguide (wslot ¼ (A) 200 and (B) 50 nm, with the waveguide width of 300 nm), and a Si wire waveguide (waveguide width ¼ 600 nm). Adapted from Tomono Y and Shimizu H (2019) CO2 Detection with Si Slot Waveguide Ring Resonators toward On-chip Specific Gas Sensing, Conference on Lasers and Electro-Optics 2019, JTh2A.96, with permission, Copyright (2019) Optical Society of America.

40

Physical Sensors: Optical Sensors

Fig. 5 An optical microscope image of the fabricated ring resonator. Adapted from Tomono Y and Shimizu H (2019) CO2 Detection with Si Slot Waveguide Ring Resonators toward On-chip Specific Gas Sensing, Conference on Lasers and Electro-Optics 2019, JTh2A.96, Copyright (2019) Optical Society of America.

Fig. 6 A cross-sectional scanning electron microscope image of the fabricated Si slot waveguide. Adapted from Tomono Y and Shimizu H (2019) CO2 Detection with Si Slot Waveguide Ring Resonators toward On-chip Specific Gas Sensing, Conference on Lasers and Electro-Optics 2019, JTh2A.96, Copyright (2019) Optical Society of America.

(A)

(B)

Fig. 7 (A) TE mode transmission spectra for the Si slot ring resonator in air. (B) Sensing characteristics of CO2 by shift of transmission spectra under dry air (blue) and CO2 atmosphere (red), near a resonance wavelength 1550 nm. Adapted from Tomono Y and Shimizu H (2019) CO2 Detection with Si Slot Waveguide Ring Resonators toward On-chip Specific Gas Sensing, Conference on Lasers and Electro-Optics 2019, JTh2A.96, Copyright (2019) Optical Society of America.

Physical Sensors: Optical Sensors

41

three approaches for detecting analytes with high sensitivity and discrimination without antigen-antibody reactions are introduced. Detection of gas molecules (H2) has been reported by using a magneto-optical SPR sensor with a catalyst for the target gas molecules (Yamane, 2021, p. SCCG01). Discrimination of analytes based on the complex refractive index (both refraction and optical absorption) by using a magneto-optical SPR sensor with angular interrogation and wavelength interrogation has been reported (Kaihara et al., 2019, p. 122003). Odor detection by analyzing the output signals of an SPR sensor with Au chips modified with multiple chemical molecules has also been reported (Brenet et al., 2018, p. 9879).

Magneto-optical SPR sensor with Pd thin film for detecting H2 gas The interaction between magneto-optical (MO) activity and SPRs has been intensively investigated theoretically and experimentally. A resonant interaction can lead to significant enhancement and modulation of the MO Kerr and Faraday effects (Safarov et al., 1994, p. 3584; Hermann et al., 2001, p. 235422; Gonzalez-Diaz et al., 2007, p. 153402). Stacked nanolayers that consist of a noble metal (NM; e.g., Au and Ag) and ferromagnetic metal (FM: e.g., Co, Fe, and Ni) under total reflection conditions represent one suitable magneto-plasmon system. The FMs experience a high absorption loss, leading to an overdamped SPR. However, a stacked FM/NM structure can be used to reduce the loss owing to the lower loss in the NM. For example, the MO Kerr effect of Au/Co/Au trilayers has been significantly enhanced by enhancing the electromagnetic field associated with the SPRs. Large MO effects are desirable for chemical and biological sensors (Temnov et al., 2010, p. 107; Gong et al., 2015, p. 191104; Sepulveda et al., 2006, p. 1085). The drastic MO response in Co/Au nanolayers improves the signal-to-noise ratio, sensitivity, and detection limit of bio-chemical sensors (Regatos et al., 2010, p. 054502; Manera et al., 2012, p. 053524). The MO polar Kerr effect (rotation of polarization at reflectance) associated with SPRs in perpendicular magnetic thin films is one of the sensing quantities that can achieve higher sensitivity. Magneto-plasmon elements consisting of a CoPt/Ag stacked nanolayer with a ZnO interface layer are optimized for MO enhancement (Yamane, 2021, SCCG01). An MO SPR sensor, which uses the polar Kerr rotation as a sensing magnitude, is developed, and by forming a Pd thin film on an MO-SPR sensor composed of a CoPt/ZnO/Ag nanolayer, hydrogen gas is detected. A Pd thin film is formed as a hydrogen detection layer, because it is known that the optical constants of Pd change due to the hydrogen reaction (von Rottkay et al., 1999, p. 408). The hydrogen reaction in the Pd layer leads to a change in the MO activity due to the shift of the SPR conditions. A Pd surface layer is expected to act as a transducer for both hydrogen detection and SPR. Fig. 8A and B show schematic illustrations of the sensor element and measurement setup for hydrogen detection. The MO responses of the stacked film in response to exposure to a H2/N2 gas mixture at room temperature are observed in an experiment. The MO–SPR sensor consisted of a stacked film of [Pd(3)/Al2O3(5)/CoPt(3.9)/Ag(9)/ZnO(30), unit nm] to which a 5 nm thick Al2O3 layer is applied as a protective layer. Fig. 9 shows the MO Kerr rotation DqK in response to exposure to a H2/N2 mixture with a concentration of 0.5–3%. DqK sharply increases after exposure to hydrogen gas and recovers after the hydrogen injection is shut off. The variation of DqK increases with increasing hydrogen concentration. Optical detection is one promising candidate technology for safe hydrogen sensing because of the electrical isolation and ability to detect at room temperature. The detection of hydrogen gas by the combination of a catalyst and a MO SPR sensor can be applied to the detection of other gasses by appropriately selecting the material.

Magneto-optical SPR sensor for discrimination of analytes based on the complex refractive index In the previous section, the MO polar Kerr effect associated with SPRs of perpendicular magnetic thin films were introduced as quantities that can achieve for higher sensitivity. The normalized reflectivity change by magnetization reversal induced by the transverse magneto-optical Kerr effect (TMOKE) has been used as quantities for higher sensitivity of the change in refractive index. As shown in the previous section, introduction of analyte bring the change of the optical absorption (Dk) as well as the change of the refractive index (Dn). The possibility of discrimination of analytes based on the difference of the complex refractive index (Dn and Dk) of the dielectric layer is reported in magneto-optical SPR sensors consisting of Al2O3/SiO2/Fe/Au structures with different Fe thicknesses, as shown in Fig. 10 (Kaihara et al., 2019, p. 122003). The quantitative sensing performance of conventional plasmonic biosensors is often evaluated by using the refractive index sensitivity, S ¼ Dq/Dn or Dl/Dn, where Dq and Dl are the change of the resonance angle and wavelength of surface plasmon, and Dn is refractive index change by introduction of analyte. The figure of merit, FoM, which is defined as the sensitivity S divided by the line width of the resonance curve (angle or wavelength dependence of the reflectivity R and normalized reflectivity change, TMOKE DR/R), g, is S/g [RIU 1], and is equivalent to v(DR/R)/v n ¼ (v(DR/R)/v q or v(DR/R)/v l z 1/g)  (v q/v n or v l/v n z S). Fig. 11 shows the minimum and maximum TMOKE (DR/R) for a sample with an Fe layer thickness of t ¼ 20 nm (A) as a function of the wavelength at fixed angles (58o and 58.1o) of incidence and (B) as a function of the angle at fixed wavelengths (1133.47 and 1136.89 nm). Fig. 11C and D show (DR/R) for the sample with t ¼ 128 nm. The change of TMOKE (DR/R) and FOM are larger for angle interrogation than that for wavelength interrogation in the sample with t ¼ 20 nm. The change of TMOKE (DR/R) and slope are larger for wavelength interrogation than that for angle interrogation in the sample with t ¼ 128 nm. FoM(l) for t ¼ 128 nm is 37 RIU 1 with respect to Dn and 370 RIU 1 with respect to Dk, showing that the FOM for Dk is larger than that for Dn, and FoM(q) z 240 RIU 1 with respect to Dn and 96 RIU 1 with respect to Dk, showing that FOM for Dn is larger than that for Dk. The FoM(l) for t ¼ 20 nm is 53 RIU 1 with respect to Dn and 730 RIU 1 with respect to Dk, showing that the FOM

42

Physical Sensors: Optical Sensors

Fig. 8 Schematic illustrations of the (A) magneto-plasmon sensor element with a Pd surface layer and (B) measurement setup for hydrogen gas detection. The change of MO Kerr rotation DqK is analyzed with a differential detector with a polarized beam splitter and two photomultiplier tubes. An external magnetic field is not applied during hydrogen detection. Adapted from Yamane H (2021) Magneto-optical surface plasmon resonances on perpendicular magnetic thin films consisting of CoPt/ZnO/Ag stacked nanolayers. Japanese Journal of Applied Physics 60: SCCG01, Copyright (2021) The Japan Society of Applied Physics.

Fig. 9 MO response DqK to the exposure to a H2/N2 mixture gas with a hydrogen concentration of 0.5–3% for MO-SPR sensor. Adapted from Yamane H (2021) Magneto-optical surface plasmon resonances on perpendicular magnetic thin films consisting of CoPt/ZnO/Ag stacked nanolayers. Japanese Journal of Applied Physics 60: SCCG01, Copyright (2021) The Japan Society of Applied Physics.

43

Physical Sensors: Optical Sensors

Fig. 10 Schematic illustrations of the magneto-optical SPR sensor consisting of Al2O3/SiO2/Fe/Au structure with the thickness of the Fe layer t on a c-sapphire substrate for discrimination of analytes based on the complex refractive index. Analyte is introduced at the section of the SiO2 layer. Adapted from Kaihara T, Shimodaira T, Suzuki S, Cebollada A, Armelles G, et al. (2019) Fe thicknesses dependence of attenuated total reflection response in magnetoplasmonic double dielectric structures: angular versus wavelength interrogation, Japanese Journal of Applied Physics 58: 122003, with permission, Copyright (2019) The Japan Society of Applied Physics.

Wavelength interrogation 1.0

Angular interrogation 1.0

(A)

t = 20 nm 0.5

ΔR/R

Δ R/R

0.5 0.0

θinc = 58.0° θinc = 58.1°

–0.5 –1.0 1000 1.0

1050

1100

1150

1200

1250

1300

–1.0 57.0 1.0

(C)

t = 128 nm

Δ R/R

Δ R/R

λ = 1136.89 nm λ = 1133.47 nm 58.0

57.5

(D)

58.5

59.0

t = 128 nm

0.5

0.0

–1.0 1100

0.0

–0.5

0.5

–0.5

t = 20 nm

(B)

–0.5

θinc = 55.3° θinc = 55.4° 1200

1300

0.0

1400

Wavelength [nm]

1500

–1.0 52

λ = 1133.47 nm λ = 1341.37 nm 53

54

55

56

57

58

Angle of incidence [°]

Fig. 11 The minimum and maximum TMOKE (O R/R) for the sample with Fe layer thickness t ¼ 20 nm (A) as a function of the wavelength at fixed angles of incidence and (B) as a function of the angle of incidence at fixed wavelengths. (C and D) The same for the sample t ¼ 128 nm. Adapted from Kaihara T, Shimodaira T, Suzuki S, Cebollada A, Armelles G, et al. (2019) Fe thicknesses dependence of attenuated total reflection response in magnetoplasmonic double dielectric structures: angular versus wavelength interrogation, Japanese Journal of Applied Physics 58: 122003, with permission, Copyright (2019) The Japan Society of Applied Physics.

for Dk is larger than that for Dn, and FoM(q) z 710 RIU 1 with respect to Dn and 520 RIU 1 with respect to Dk, showing that the FOM for Dn is larger than that for Dk. Since the angle and wavelength interrogation for DR/R are respectively related to the real and imaginary parts of the propagation constant of SPPs in magneto-optical SPR sensors, it is possible to detect the change of the complex refractive index of an analyte. When an analyte is introduced into the SiO2 layer as a refractive index deviation from the original index, nSiO2 (real number), as nSiO2 þ Dn þ iDk, the deviation in the real part, Dn, and the imaginary part, Dk, are reflected in the shift of DR/R

44

Physical Sensors: Optical Sensors

along the angle axis and the wavelength axis, respectively. Theoretical shifts in the angular interrogation and the wavelength interrogation modes are shown in Fig. 12. The red curves represent the case in which Dn is variable as a function of the horizontal axis while Dk is kept at 0, and the blue curves represent the case in which Dk is variable as a function of the horizontal axis while Dn is kept at 0 at t ¼ 128 nm. Although either Dn or Dk causes a shift on both the angle and the wavelength axes, Dn affects the shift in the angle direction dominantly over Dk (Fig. 12A). On the other hand, Dk causes a shift in the wavelength more largely than Dn (Fig. 12B). Since the Dk causes negative shifts in the angle direction, the absolute values are plotted for comparison in Fig. 12A. Since the complex refractive index (combination of refraction and optical absorption) is unique to the analyte, these results suggest the feasibility of the two interrogation modes as additional information on analyte discrimination as well as information of the concentration. The use of shorter wavelength (ultraviolet to far ultraviolet light) is effective for sensing based on complex refractive index with higher sensitivity, because the optical absorption of molecules is more enhanced in the wavelength of 120– 400 nm, compared with that of visible wavelength (400–700 nm).

VOC sensor with SPR imaging with cross-reactive sensor microarray Surface plasmon resonance imaging (SPRI) by combining the strength of SPR with cross-reactive microarrays is developed in order to monitor volatile organic compounds (VOCs) in highly-selective optoelectronic nose (eN) applications (Brenet et al., 2018, p. 9879). SPRI is promising for eN development, because it is possible to immobilize up to hundreds of sensing molecules on the same chip for the creation of a large sensor array. The number of sensors is limited by the resolution of the microarray printing of the sensing molecules. The interactions between the analytes and hundreds of sensors can be simultaneously monitored in an image using the same instrument. SPRI allows an odor to be recorded in the form of an image, based on which a characteristic pattern can be generated for each sample. This information can be used for the discrimination of samples with pattern recognition and multivariate statistics. Temporal responses with kinetic information can be obtained. SPRI microarrays containing cross-reactive sensors are used as transducers, consisting of a prism coated with a thin Au layer. On the Au surface, sensing molecules, which are monolayer of 18 biomimetic peptides and organic molecules (Brenet et al., 2018, p. 9879), are deposited in a microarray format, based on the studies of short peptides screened by molecular docking simulations (Mascini et al., 2017, p. 161). An SPRI system for gas sensing is based on the Kretschmann configuration, as shown in Fig. 13A (Brenet et al., 2018, p. 9879). A collimated beam with a wavelength of 632 nm from an LED is polarized and sent toward the functionalized Au surface through the prism to illuminate the entire microarray. SPR causes a progressive variation in the reflectivity (%R), for example, as a function of the incident angle of light, as shown in Fig. 13B. When VOCs present in the dielectric medium in the gas phase interact with the sensing material on the microarray, there is an alteration in the local refractive index. By choosing an angle of incidence at the highest slope of the plasmon curves, small changes in the resonance conditions can produce large variations in the reflectivity. Upon interaction between VOCs and all the

Shift in angle [°]

1

(A)

0.100

'n

0.010 0.001

'k

10–4 10–5

Shift in wavelength [nm]

10–6 100 10

10–5

10–4

0.001

0.010

0.001

0.010

(B) 'k

1 0.100

'n

0.010 0.001 10–6

10–5

10–4

Deviation 'n,'k [RIU] Fig. 12 The shifts in the angular interrogation (A) and the wavelength interrogation (B) modes caused by the refractive deviation of the real, Dn, and imaginary, Dk, components from the refractive index of SiO2. Adapted from Kaihara T, Shimodaira T, Suzuki S, Cebollada A, Armelles G, et al. (2019) Fe thicknesses dependence of attenuated total reflection response in magnetoplasmonic double dielectric structures: angular versus wavelength interrogation, Japanese Journal of Applied Physics 58: 122003, with permission, Copyright (2019) The Japan Society of Applied Physics.

Physical Sensors: Optical Sensors

Gas outlet

(A)

(B)

(C)

25°C

LED 632 nm

(E)

Tw

Lens

'%R

%R

Lens Polarizer

Chip

CCD camera

(D)

Time

Incident angle

P

VOC analyte line

'%R(t)

Gas inlet

45

F

F

Outlet + VOCs exit PID control

SPRI analysis cell

Reference/Dilution line Fig. 13 Schematic illustration of the setup of the optoelectronic nose. (A) Configuration of the SPRI for the analysis of VOCs. (B) Plasmon curves for a given sensor before (black) and after (red) the interaction with a VOC. (C) Sensorgrams: a set of kinetic binding curves obtained with three sensors, plotted with the variation of reflectivity D%R(t) as a function of time upon the injection of a VOC at a fixed working angle qw; (D) SPRI differential image of the microarray recorded by the CCD camera after exposure to a VOC sample, and (E) the fluid bench constructed for gas generation, sampling and regeneration of the optoelectronic nose. Reprinted (adapted) with permission from Brenet S, John-Herpin A, Gallat FX, Musnier B, Buhot A, et al. (2018) Highly-selective optoelectronic nose based on surface plasmon resonance imaging for sensing volatile organic compounds. Analytical Chemistry 90: 9879–9887. Copyright {2018} American Chemical Society.

sensors on the microarray, the shift of the plasmon curves induces reflectivity changes D%R as shown in Fig. 13B and C. Fig. 13D shows an SPR image, where the spots light up with different intensities related to interaction affinities. VOCs from diverse chemical families having different properties and distinct smells (alcohols, esters, carboxylic acids, ketones, hydrocarbons, aldehydes, and amines) are tested in order to evaluate the efficiency of the optoelectronic nose based on SPRI. Fig. 14A–F show sensorgrams of six VOCs, where D%R(t) is the baseline-corrected reflectivity change upon exposure of the sensor array to the analytes (Brenet et al., 2018, p. 9879). All the sensorgrams have characteristic shapes with an association phase, equilibrium, and dissociation phase, showing that SPRI is very efficient for sensing VOCs in the gas phase. The adsorbed VOCs can be refreshed by clean air. For every VOC, all the cross-reactive sensors give responses with different affinities, with the exception of the internal reference sensors S0 that do not interact with any VOC and are thus used for a negative control. These results confirm that the 18 organic molecules and peptides from the microarray are relevant and effective. Among these VOCs’ sensorgrams, the set of cross-reactive sensors responds very differently to them with distinct kinetic interactions, even for the VOCs of the same family, such as 1octanol and 1-pentanol. To illustrate the array’s capabilities to discriminate between VOCs, SPRI images at equilibrium (10 min after VOC injection) are given, where the sensor array lights up with different gray levels depending on affinities between the sensors and VOCs. Moreover, a pattern for each VOC is generated by plotting the overall sensor response at equilibrium (Fig. 14A0 –F0 ). Based on SPRI images and the response profiles at equilibrium, it is possible to distinguish between certain VOCs, for instance between phenol and 2-methylpyrazine. The performance of the system is evaluated in terms of sensitivity, repeatability, stability, and reusability. To determine the limit of detection (LOD), calibration curves are established using VOCs with different volatilities (ethanol and 1-octanol). The optoelectronic nose is successively exposed to 1-octanol at concentrations ranging from 1.43 to 5.84 ppm, and then to ethanol at a higher concentration range (from 170 to 1400 ppm). Fig. 15A shows the real-time response of different sensors upon exposure to 1octanol. For easier visualization, only results on the internal reference and two randomly selected sensors are given as examples. Fig. 15B shows the calibration curves for three randomly selected sensors and the internal reference. The variation of reflectivity is proportional and almost linear. Using their slope, the sensitivity is between 0.2 and 0.4 D%R$ppm 1 for all the sensors. High selectivity and discrimination between VOCs differing by a single carbon atom are reported on the system. A VOC sensor system with SPR imaging with a cross-reactive sensor microarray is applied to Si-waveguide-based RI sensors (https://aryballe.com/).

Summary The principles, classification, and recent advances in optical sensors for application to biosensors, and gas sensors are described. Advances in optical equipment, integrated optics based on free space and waveguide optics, designs based on novel approaches

(A)

(B)

(C)

(D)

(E)

(F)

(A’)

(B’)

(C’)

(D’)

(E’)

(F’)

Fig. 14 Array responses, characteristic SPR images, and response patterns for selected VOCs. Sensorgrams depicting the sensor array response after exposure to (A) 2-methylpyrazine (290 ppm), (B) phenol (34 ppm), (C) isoamyl butyrate (70 ppm), (D) 1-pentanoic acid (51 ppm), (E) 1pentanol (47 ppm), and (F) 1-octanol (7.6 ppm). The chemical structure for each VOC is inserted in the inset. SPR images and response patterns at equilibrium for (A0 ) 2-methypyrazine, (B0 ) phenol, (C0 ) isoamyl butyrate, (D0 ) 1-pentanoic acid, (E0 ) 1-pentanol, and (F0 ) 1-octanol. Reprinted (adapted) with permission from Brenet S, John-Herpin A, Gallat FX, Musnier B, Buhot A, et al. (2018) Highly-selective optoelectronic nose based on surface plasmon resonance imaging for sensing volatile organic compounds. Analytical Chemistry 90: 9879–9887. Copyright {2018} American Chemical Society.

Physical Sensors: Optical Sensors (A)

VOC measurement (ppm)

54 1.43

52 %R (t)

47

2.09

1.68

3.03

4.01

5.84

50 48 46 44 0

(B)

3

50

75 100 125 Time (min)

150

175

S0

2,5

S7

2 '%R

25

S8 S18

1,5 1 0,5 0 0

1

2 5 3 4 VOC concentration (ppm)

6

Fig. 15 Sensitivity and limit of detection of the optoelectronic nose: (A) the variation of raw reflectivity as a function of time upon exposure to various concentrations of 1-octanol, for the internal reference sensor (black) and two randomly selected sensors (S7 in gray and S8 in red), (B) calibration curves: averaged D%R at equilibrium for the four replicates of the internal reference sensor (black) and the three randomly selected sensors S7, S8, and S18 (gray, red, and green, respectively) as a function of 1-octanol concentration and the associated linear regression curve. Reprinted (adapted) with permission from Brenet S, John-Herpin A, Gallat FX, Musnier B, Buhot A, et al. (2018) Highly-selective optoelectronic nose based on surface plasmon resonance imaging for sensing volatile organic compounds. Analytical Chemistry 90: 9879–9887. Copyright {2018} American Chemical Society.

such as synthesis of dissimilar materials and signal processing in combination with materials informatics, or artificial intelligence will pave the way toward highly sensitive novel biosensors and gas sensors with analyte discrimination. Especially, owing to recent development of small-size light sources and detectors of infrared light (Giorgetta et al. 2009, p. 1039; Gmachl et al., 2021, p. 1533), NDIR optical sensors will be developed and available at lower price. Although combination of optical sensors based on waveguide optics and use of light with shorter wavelength (ultraviolet to far ultraviolet light) has not been fully developed yet, recent development of light emitting devices for blue and ultraviolet light (Takano et al., 2017, p. 031002) will enable sensing based on complex refractive index with higher sensitivity.

References Barrios, C.A., 2009. Optical slot-waveguide based biochemical sensors. Sensors 9, 4751–4765. Brenet, S., John-Herpin, A., Gallat, F.X., Musnier, B., Buhot, A., et al., 2018. Highly-selective optoelectronic nose based on surface plasmon resonance imaging for sensing volatile organic compounds. Analytical Chemistry 90, 9879–9887. Claes, T., Molera, J.G., Vos, K.D., Schacht, E., et al., 2009. Label-free biosensing with a slot-waveguide-based ring resonator in silicon on insulator. IEEE Photonics Journal 1, 1943. Giorgetta, F.R., Baumann, E., Graf, M., Yang, Q., et al., 2009. Quantum cascade detectors. IEEE Journal of Quantum Electronics 45, 1039–1052. Gmachl, C., Capasso, F., Sivco, D.L., Cho, A.Y., 2021. Recent progress in quantum cascade lasers and applications. Reports on Progress in Physics 64, 1533. Gong, Y., Li, K., Caver, S., Martinez, J.J., Huang, J., et al., 2015. Current control of light by nonreciprocal magnetoplasmonics. Applied Physics Letters 106, 191104. Gonzalez-Diaz, J.B., Garcia-Martin, A., Armelles, G., Garcia-Martin, J.M., Clavero, C., et al., 2007. Surface-magnetoplasmon nonreciprocity effects in noble-metal/ferromagnetic heterostructures. Physical Review B 76, 153402. Hermann, C., Kosobukin, V.A., Lampel, G., Peretti, J., Safarov, V.I., et al., 2001. Surface-enhanced magneto-optics in metallic multilayer films. Physical Review B 64, 235422. Hodgkinson, J., Smith, R., Ho, W.O., Saffell, J.R., Tatam, R.P., 2013. Non-dispersive infra-red (NDIR) measurement of carbon dioxide at 4.2 mm in a compact and optically efficient sensor. Sensors and Actuators B: Chemical 186, 580. Ito, H., Nogata, Y., Matsuzaki, S., Kuboyama, A., 1969. Vacuum-ultraviolet absorption spectra of aliphatic ketones. Bulletin of the Chemical Society of Japan 94, 2453–2458. Kaihara, T., Shimizu, H., 2017. Nonreciprocal dielectric-loaded plasmonic waveguides using magneto-optical effect of Fe. Optics Express 25, 730. Kaihara, T., Shimodaira, T., Suzuki, S., Cebollada, A., Armelles, G., et al., 2019. Fe thicknesses dependence of attenuated total reflection response in magnetoplasmonic double dielectric structures: Angular versus wavelength interrogation. Japanese Journal of Applied Physics 58, 122003. Manera, M.G., Colombelli, A., Rella, R., Caricato, A., Cozzoli, P.D., et al., 2012. TiO2 brookite nanostructured thin layer on magneto-optical surface plasmon resonance transductor for gas sensing applications. Journal of Applied Physics 112, 053524. Mascini, M., Pizzoni, D., Perez, G., Chiarappa, E., di Natale, C., et al., 2017. Tailoring gas sensor arrays via the design of short peptides sequences as binding elements. Biosensors and Bioelectronics 93, 161–169. National Astronomical Observatory of Japan, 2019. Chronological Scientific Tables. Maruzen.

48

Physical Sensors: Optical Sensors

Nobre, M., Fernandes, A., Ferreira da Silva, F., Antunes, R., et al., 2008. The VUV electronic spectroscopy of acetone studied by synchrotron radiation. Physical Chemistry Chemical Physics 10, 550–560. Pope, R.M., Fry, E.S., 1997. Absorption spectrum (380–700 nm) of pure water. II. Integrating cavity measurements. Applied Optics 36, 8710–8723. Regatos, D., Farina, D., Calle, A., Cebollada, A., Sepulveda, B., et al., 2010. Au/Fe/Au multilayer transducers for magneto-optic surface plasmon resonance sensing. Journal of Applied Physics 108, 054502. Robinson, J.T., Chen, L., Lipson, M., 2008. On-chip gas detection in silicon optical microcavities. Optics Express 16, 4296. Safarov, V.I., Kosobukin, V.A., Hermann, C., Lampel, G., Peretti, J., et al., 1994. Magneto-optical effects enhanced by surface plasmons in metallic multilayer films. Physical Review Letters 73, 3584. Samoc, A., Miniewicz, A., Samoc, M., Grote, J.G., 2007. Refractive-index anisotropy and optical dispersion in films of deoxyribonucleic acid. Journal of Applied Polymer Science 105, 236–245. Sepulveda, B., Calle, A., Lechuga, L.M., Armelles, G., 2006. Highly sensitive detection of biomolecules with the magneto-optic surface-plasmon-resonance sensor. Optics Letters 31, 1085. Shimizu, H., Noriyasu, H., 2014. Measurement of carbon dioxide concentration by fiber-loop ring-down spectroscopy for continuous remote measurement. Japanese Journal of Applied Physics 53, 116601. Shimizu, H., Ogura, T., Maeda, T., Suzuki, S., 2021. A wedge-shaped Au thin film: Integrating multiple surface plasmon resonance sensors in a single chip and enhancing the figure of merit. Frontiers in Nanotechnology 3, 724528. Takano, T., Mino, T., Sakai, J., Noguchi, N., et al., 2017. Deep-ultraviolet light-emitting diodes with external quantum efficiency higher than 20% at 275 nm achieved by improving light-extraction efficiency. Applied Physics Express 10, 031002. Tarsa, P.B., Rabinowitz, P., Lehmann, K.K., 2004. Evanescent field absorption in a passive optical fiber resonator using continuous-wave cavity ring-down spectroscopy. Chemical Physics Letters 383, 297. Temnov, V.V., Armelles, G., Woggon, U., Guzatov, D., Cebollada, A., et al., 2010. Active magneto-plasmonics in hybrid metal–ferromagnet structures. Nature Photonics 4, 107. Tomono, Y., Shimizu, H., 2019. CO2 detection with Si slot waveguide ring resonators toward on-chip specific gas sensing. In: Proceeding of Conference on Lasers and Electro-Optics 2019. JTh2A.96. Tong, Z., Wright, A., McCormick, T., Li, R., Oleschuk, R.D., Loock, H.-S., 2004. Phase-shift Fiber-loop ring-down spectroscopy. Analytical Chemistry 76, 6594. von Rottkay, K., Rubin, M., Duine, P.A., 1999. Refractive index changes of Pd-coated magnesium lanthanide switchable mirrors upon hydrogen insertion. Journal of Applied Physics 85, 408. Xu, Y., Bai, P., Zhou, X., Akimov, Y., Png, C.E., et al., 2019. Optical refractive index sensors with plasmonic and photonic structures: Promising and inconvenient truth. Advanced Optical Materials 7, 1801433. Yamane, H., 2021. Magneto-optical surface plasmon resonances on perpendicular magnetic thin films consisting of CoPt/ZnO/Ag stacked nanolayers. Japanese Journal of Applied Physics 60, SCCG01. Yariv, A., Yeh, P., 2006. Photonics: Optical Electronics in Modern Communications, 6th edn. In: The Oxford Series in Electrical and Computer Engineering. Oxford University Press, Oxford. Ye, J., Shioi, M., Lodewijks, K., Lagae, L., Kawamura, T., 2010. Tuning plasmonic interaction between gold nanorings and a gold film for surface enhanced Raman scattering. Applied Physics Letters 97, 163106. Yeh, Y.L., 2008. Real-time measurement of glucose concentration and average refractive index using a laser interferometer. Optics and Lasers in Engineering 46, 666–670. Zhang, H.Q., Boussaad, S., Tao, N.J., 2003. High-performance differential surface plasmon resonance sensor using quadrant cell photodetector. The Review of Scientific Instruments 74, 150–153.

Relevant websites http://www.cfa.harvard.edu/HITRAN/dThe HITRAN Database. https://www.yokogawa.com/us/solutions/products-platforms/process-analyzers/gas-analyzers/ftnir-ir/dWebsite of YOKOGAWA Corporation of America. https://www.picarro.com/company/technology/crdsdWebsite of Picarro Inc. https://aryballe.com/dWebsite of Aryballe technologies.

Physical Sensors: Plasmonic Sensors Yuzuru Iwasaki, NTT Device Technology labs., NTT Corporation, Atsugi, Japan © 2023 Elsevier Ltd. All rights reserved.

Introduction Refractive index and biosensing Principle of RI measurement by SPR Planar SPR Prism materials Sensor plate detachable from prism Light source Otto configuration SPREETA Localized SPR (LSPR) Chemically synthesized LSPR particles Nanosphere lithography for LSPR devices Bottom-up fabrication for LSPR devices Imaging SPR One-dimensional SPR Fiber/waveguide based SPR Portable SPR Selectivity and sensitivity of SPR sensors Selectivity Combination of SPR and Raman spectroscopy Sensing volume, sensitivity of SPR Noise source Mass transport Data processing in SPR Comparison to other methods Future perspective References

50 50 51 52 53 53 53 54 54 54 54 54 55 55 56 56 57 57 57 57 57 58 58 58 59 59 59

Glossary Evanescent wave Non propagating light, confined on the total reflection plane. The wavenumber of the evanescent light can be smaller than that of normal propagating light. Molecular recognition A property of a biological macromolecule that reacts with a specific molecule. For example, an antibody binds exclusively with its antigen, an enzyme catalyzes the reaction with a specific substrate molecule, and single stranded DNAs forms a double strand only with a complementary DNA sequence. The antibody, enzyme, and DNA are called molecular recognition molecules. Optical fiber A transparent thin string of glass that transmits light through its core (center). The core has a higher refractive index than the surrounding cladding; therefore, the light is confined in the center of the fiber for a long distance. Optical fiber is used mainly in telecommunications. For a sensor, an exposed core is used as the sensing part. Optical waveguide An optical device with a planar plate that embeds an optical fiber structure on the plane. An optical waveguide can integrate optical elements, such as a light source/detector, mirror, lens, Y-splitter/demultiplexer, phase shifter, and is configured as an interferometer, or wavelength selector. In a biosensor, the interferometer is used as the sensing parts. P polarization When a light is reflected by a flat surface, the direction of the electric field of light that induces electron movement on the surface out of the plane interacts strongly with the surface material. The incidence plane includes the normal vector of the reflecting plane and the direction vector of the light. P polarization is also called the transverse magnetic (TM) mode. P-polarization light is generated or separated from a natural light passing through a polarizer with an appropriate orientation angle. Refractive index A complex number describing optical properties such as reflectivity, transmittance, and absorbance of the light path through the material. In a biosensor, the refractive index is a measure of mass of the analyte molecules on the sensing surface.

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00050-9

49

50

Physical Sensors: Plasmonic Sensors

Abstract A plasmon sensor is a type of optical transducer for a biosensor. In particular, surface plasmon resonance (SPR) is widely used in biological analysis. Plasmon sensors transform the biological reaction of molecular recognition materials into a refractive index, enhanced photon emission and adsorption. This article describes the optical configurations and detection principles of SPR devices and discusses their characteristics, sensitivity, sensor data processing.

Introduction Surface plasmon resonance (SPR) is an excitation phenomenon of free electrons on a metal surface by light, and mainly used to measure local refractive indexes (RIs) in biosensors. In the context of biosensor, the term SPR also refers to the methodology. With the SPR, changes in the RI can be measured on a sensor’s surface in real time and in situ. Standard biological methods require pretreatment of sample like a fluorescent labeling, which potentially interferes with the native molecular interactions. SPR eliminates the need for labeling molecules (Masson, 2017). In addition, the degree of capturing and dissociation can be measured quantitatively as function of time, and kinetic information of binding reactions can be measured. Another major advantage of the SPR method is its capability of multiplexing detection or imaging format measurement. Many types of multi point detection devices and microscopic imaging techniques have been developed and applied for array detection and imaging of naïve cellular functioning. Nevertheless, the optical configurations for SPRs are rather simple in that, basically they are internal reflection spectroscopy. In the 1980s, researchers started using SPR to determine conjugation parameters of two interacting molecules,done immobilized on the sensor surface, and the other flowing over the sensor surface. A sophisticated instrument Biacore was commercialized in this field. Then, with the advent of nano-structure fabrication techniques, localized surface plasmon resonance (LSPR) devices were developed, which utilize nano scale particles and patterns as plasmon media. Comprehensive reviews have been continuously published (Homola et al., 1999; Homola, 2003; Fan et al., 2008; Choi and Choi, 2011; Sharma et al., 2018), and a Book focusing on SPR (Schasfoort, 2017) is available. Moreover, in recent years, SPR has been used not only for high-end biological interaction analysis but also for on-site and easeof-use biological measurement. The use of smartphones in conjunction with SPR methods is notable in application fields. The sensitivity for the smallest RI change is one of the quantitative performance measures of an SPR sensor. The effective sensitivity to RI changes has been continuously improved throughout the development history, though the principal sensitivity for RI change is determined by the optical characteristics of the plasma medium metal. This entry covers SPR technology for biosensors including how the RI helps biosensing, the physical principles of SPR measurement, the types of SPR devises, and the performance of SPR devices and recent research and developments of emerging techniques.

Refractive index and biosensing SPR methods can detect small changes in the RI of biological sample. The RI can be used as a measure of mass, which could be changed by antigen-antibody reactions, DNA double-strand formation, and any molecules having affinity. Let’s consider the case of an immobilized antibody and a solved antigen in solution (Fig. 1). An antigen and its antibody bind and form a specific and strong antigen-antibody complex by multiple hydrogen bonds. Before the antigen is conjugated with antibody, the local RI in the vicinity of the antibody is the average of the immobilized antibody and surrounding solution molecules. Then the antigen comes into contact with the immobilized antibody and forms an antigen-antibody complex. The volume previously occupied

Antigen Solution

Antibody

Substrate (gold)

Beforeconju gat ion Fig. 1

RI change caused by antigen-antibody reaction.

Substrate (gold)

Aft er conjugation

Physical Sensors: Plasmonic Sensors

51

by the solution molecules (dashed circle) is replaced with the antigen. After the complex formation, the local RI of the surface changes to the volume fraction average of the antigen, antibody, and solution molecules. SPR devices detect this difference in the local RI and quantitate the conjugate formation. Because their sensitivity is localized within reach of the surface plasmon, which is confined to less than the wave length scale of the excitation light, the molecular size difference of the average RI can be measured.

Principle of RI measurement by SPR To observe SPR we need an optical setup like the one shown in Fig. 2A. P-polarized [transverse magnetic (TM) mode] incident light is reflected at the prism surface by total internal reflection and an evanescent wave is generated outside of the prism at reflection point. The wave excites surface plasmon in metal, resulting in dispersion the light energy as heat, and thus the reflected light is dimmed. This dimmed reflection is a characteristic of SPR. In the resonance condition, determined by the wavelength and incident angle of the light, the amplitude of electric field is maximized at the gold-sample interface, and then sample’s dipole moment is interacted by the amplified electric field. Therefore, the resonance condition shifts depending on the dielectric constant within reach of evanescent wave. On the contrary, we can detect small RI changes on the gold surface from the changes in the reflection profile. SPR devices are classified by their plasma medium structure, which is either (planar) SPR (Fig. 2A, B, and D) or localized SPR (LSPR, Fig. 2C). SPR signal can be found in the incident angle, incident wavelength, and reflection intensity. These are called interrogation modes. SPR technology is also classified by its functions, which are mainly quantitative RI measurement (Fig. 2A–C) and RI imaging (Fig. 2D). SPR and LSPR can provide both functions.

(A)

(B)

(C) (D)

Fig. 2 (A) Kretschmann configuration of SPR optics and layer structure of sample. (B) Otto SPR optical configuration. (C) Optical setups of LSPR devices. (D) Optical configuration of SPR imaging.

52

Physical Sensors: Plasmonic Sensors

Planar SPR Several optical configurations can attain the SPR condition. The most common one is called the Kretschmann configuration, which uses planar reflection surfaces as shown in Fig. 2A. In this configuration, a gold thin film (thickness  50 nm) is attached to a high RI glass prism, and a sample is on the gold film. Another, less common one is called the Otto configuration (described later). In the Kretschmann configuration, the gold is illuminated from the glass side, and surface plasmon is excited. The light does not pass through the sample layer in this configuration. The p-polarized reflectance (Rp) dependence on the incident angle or incident wavelength shows a valley at incident angles above the critical angle of prism (BK7) against n4 (water). In Fig. 3A, the Rp is shown as a pseudo color as a function of incident angle and wavelength, and the white curves show the Rp at 770 nm and an incident angle of 75 degrees. These characteristic curves are called SPR curves, and the incident angle at the bottom of the is called SPR angle. In Fig. 3A, the dark blue part corresponds to reflection damped by SPR. When the RI of the gold surface increases, the incident angle-reflectance curve shifts to a higher incident angle or longer wavelength. This shift depends on the RI change on the gold surface, which is why we can use SPR for the biosensors. The SPR curve can be calculated using the Fresnel multi-layer reflection formula (Kessler and Hall, 1996). This technology is the part of design process for dielectric multilayer optical coatings, and commercial software in this field can calculate SPR curves with an RI database. In the Fresnel reflection model, all components of in the optical path are considered to be stacked as uniform thin films each having specific complex RI and thickness (except for the first (prism) and last layer, whose thickness are infinite). As a simple example, let’s think of a four-layer stack comprising a prism (RI ¼ n1; thickness infinite), a gold metal film (RI ¼ n2; thickness d2), a sample layer (RI ¼ n3; thickness d3), and a solution layer (RI ¼ n4; thickness infinite). The dependence of the SPR curve on the sample’s RI is calculated as shown in Fig. 3B. By determining the SPR angle from the measured SPR curve, we can obtain the RI n3. This is called the incident angle interrogation mode. Fig. 3C shows the relation between the SPR angle and n3 when d3 is 200 nm.

(A)

1 Rp min bySPR

p-Reflectance

Incident angle [degree]

80

70

60 400

600

800

900

0

Wavelength [nm] (B)

(C)

Fig. 3 (A) Rp dependence on the incident angle and wavelength, calculated by Fresnel reflection mode of n1 ¼ 1.51, n2 ¼ 0.18 þ 4.47i, d2 ¼ 50 nm, n3 ¼ 1.4, d3 ¼ 200 nm, n4 ¼ 1.33. The white graphs are SPR curves for wavelength of 770 nm and for incident angle of 75 degrees. SPR curves for 4 layer model. (B) Calculated SPR curves of different layer 3 RI (n3) using Fresnel multilayer formula. (C) The relation of SPR angle and layer 3 RI (n3). The layer structure is shown in Fig. 2A.

Physical Sensors: Plasmonic Sensors

53

For biosensing, one of the interacting molecular pairs is immobilized on the gold surface. This immobilization layer can be layer 3. A more complex sensor layer structure can be modeled by specifying the thickness and complex RI for each layer. Note that even the d3 is as thin as the size of biological macromolecules such as an antibody, we can measure the n3 by SPR. The Rp at the SPR angle is controlled by the gold thickness. By using a monochromatic light source, it can be optimized to nearly zero. Gold thin film is commonly used for biosensors for its stability in electrolyte and thiol-group chemistry for immobilization of biomolecules. Other metals, silver, copper, and aluminum (Tanabe et al., 2017), also show SPR in the near-infrareddultra-violet wavelength region; however, they are less commonly used in biosensor. SPR can also be measured by spectroscopic methods. Because of the dispersion of the complex RI of the surface plasmon medium (gold), the SPR curve can be obtained in the wavelength region (Lucarini et al., 2005). In this case, the incident angle is fixed and spectrum of Rp is measured. When the optical configuration in Fig. 2A is used, a dip appears in the spectrum. This is called wavelength interrogation mode. The minimum wavelength of reflection depends on the sample’s RI n3. This spectrum can also be calculated by Fresnel reflection modeling using the wavelength dispersion of the complex RI of gold (n2). This wavelength interrogation mode is common types based on optical fiber and waveguides. Most commercial instruments for biological molecular interaction measurements are based on the Kretschmann configuration.

Prism materials In the Kretschmann configuration, the SPR angle is observed when the RI of the sample is smaller than the RI of the substrate (prism). Since the larger glass RI is, the smaller the incident angle observed in SPR, it is preferable to use high RI material for the gold substrate. BK7 glass (RI:  1.52) a standard microscopy slide glass, is commonly employed because of its availability and low cost and for its comparably good flatness and optical homogeneity. When a BK7 prism is used, a bulk sample RI lower than  1.45 can measured; however, the incident angle must be more than 80 degrees, which makes it difficult to focus the incident light on the sample position. To measure higher RIs or to use smaller incident angles, SF10, SF11 glass is used. The gold film is formed on the glass material by sputtering or physical vapor deposition. To improve the adhesion of gold, chromium (Cr) or titanium (Ti) is used as the next layer on the glass. Frequently, chromium has been used, though Ti may be selected when the sample solution etches chromium or SPR measurements are conducted with electrochemistry (Iwasaki et al., 1997; Iwasaki et al., 1999, 2001, 2006). The thicker the Cr or Ti layer is, the stronger the Au binding, but with a thicker layer, the SPR curve becomes ambiguous (a shallower dip). Depending on the application, optimization of the thickness of the adhesion layer may be necessary. For low cost sensor devices, plastic prism can be used (Gawedzinski et al., 2017; Hinman et al., 2017; Lambert et al., 2018; Nootchanat et al., 2019; Walter et al., 2020; Lertvachirapaiboon et al., 2021).

Sensor plate detachable from prism The gold thin film can be formed directly on a high RI prism (a hemicylinder or dove) or on a glass plate with the same RI as the prism. For biosensors, the gold surface is chemically modified with molecular recognition molecules and blocking (to prevent nonspecific binding) layers. These layers must be as thin as the extent of evanescent wave, and flat so that the incident light is not scattered. Furthermore, for multiple analytes, different recognition molecules should be spotted using addressable spotters and ink-jet printers (Inoue et al., 2016; Shrestha et al., 2020). In carrying out these treatments for the SPR sensor surface, a flat plate sensor chip or a cartridge containing the functional flat plate is advantageous. This also enables microfluidics for sample liquid handling can be attached on the flat surface. Therefore, in most commercial SPR instruments and laboratory experimental protocols, the prism is installed in the instruments, and the sensor chip is attached to the prism after surface modification is completed. In this case, an optical adhesion is needed between the prism and the sensor chip using RI matching material so that the incident light is not reflected at the prism-sensor chip interface before reaching the glass-gold interface. The RI matching materials may be a liquid or gel. For reproducibility and convenience of replacement, gel-type RI matching material is used, because liquid type RI material is must be removed and the prism cleaned every time the sensor chip is replaced.

Light source To excite surface plasmons, a light source is needed, and the one use depend on the optical configuration. A gas laser, solid state laser, or LED (light emitting diode) can be used as a narrow band light source for intensity detection. For the wavelength interrogation, a wide-band light source is used, such as a stabilized tungsten light (Pan et al., 2018), Xenon lamp, or white LED. A monochromatic light source will yield the sharpest SPR curve. The zero-reflection dip level will be observed only when a monochromatic light source is used. The SPR curve of a polychromatic light source is considered as that of the weighted spectral average for a monochromatic light source. When a polychromatic light source is used, the SPR dip becomes wider and the reflection minimum level become shallower, but the SPR angle dependence on the RI of sample does not change. In practice, when coherent light such as a laser is used as incident light, the reflected light shows interfering pattern caused by the RI boundary in the SPR optics. Any dust on the total reflection surface causes a complex and unstable interfering pattern. Because this interfering pattern is fixed at the incident angle and convoluted with the theoretical SPR curve, RI determination from the experimental SPR curve becomes noisy.

54

Physical Sensors: Plasmonic Sensors

Otto configuration SPR can be observed in a prism/sample/gold layer system (Fig. 2B). The gold surface is separated from the prism surface and the sample layer is between the prism and gold. This is called the Otto configuration. In this configuration, since the sample is in the sub-micro-meter gap between the prism and gold, one needs to introduce an aqueous sample into this nano-gap. Different from the Kretschmann configuration, the gold does not need to be a thickness-controlled thin film, and the substrate for the gold can be any (opaque) material. The Otto configuration is quite rare in biosensing applications, though an SPR sensor using gold-deposited opaque paper in the Otto configuration has been reported (Fukuda et al., 2017).

SPREETA Texas Instruments developed a special SPR device called SPREETA (Chinowsky et al., 2003). This solid-state integrated circuit device contains a light source, a plasmon excitation surface, incident angle interrogation Kretschmann optics, and interfacing electronic circuits. The electrical connection of the devices is compatible with a dual in line (DIP) package. The commercial device is provided with an evaluation. To eliminate the needs for a lens for incident angle scanning, SPREETA utilizes unique optical path system. On the other hand, different from a point reflection system in Fig. 2A, the reflection occurs along a line, with each point corresponding to a certain incident angle.

Localized SPR (LSPR) SPR can be also observed using gold nano particles (colloids). This type of SPR is called localized SPR (LSPR) (Kravets et al., 2018). Gold and other noble metal particles less than few hundred nanometer in diameter show strong plasmon resonance. The transmission spectra of LSPR particles suspended in aqueous solution show SPR curves, and the RI of the space surrounding LSPR particle affects the resonance wavelength. The conduction band electrons in the gold nano particles directly couple with light, so LSPR can be detected with standard optical absorbance instruments (Fig. 2C) without a special coupling structure like a prism in planar SPR. The strong adsorption itself is used as a visible marker in immunochromatography (Bosch et al., 2017). The resonance wavelength is tuned by the particles’ size and shape. For example, gold nano particles dispersed in different solvent with different RIs show RIdependent shift of the adsorption band (Ghosh et al., 2004). The electric field localization of LSPR is more compact than that of SPR based on planar gold film, for examples, the RI sensitive volume is confined 29 nm from the particles’ surface (Xu and Käll, 2002). LSPR can be observed in chemically synthesized particles or in lithographically prepared nano size patterns.

Chemically synthesized LSPR particles Chemically, nano particles can be synthesized by reduction of HAuCl4. The size and shape of the nano-particles can be controlled by optimizing the reaction condition (pH, temperature, solvent, surfactants). The RI sensitivity in the shift of the resonance wavelength is governed by the aspect ratio of the nano particles. Refs. Maier et al. (2001) and Vigderman and Zubarev (2013) describe a reliable procedure for rod-like nano particle fabrication. The resonance wavelength of the nano rods can be controlled by their aspect ratio (shorter axis/longer axis). The aspect ratio of  7 resulted in a resonance wavelength of infrared bad. Ref. He et al. (2020) reported morphological control of gold nano rods. Fig. 4 shows the shape, color change and spectral shift of the gold nano-rods. Gold nano-rods are prepared by growing nano-particles in the presence of different concentration of a growth mediator (glutathione). Higher mediator concentrations produce tip structures both ends of the rod, and these structures give longer red shifts per RI change. Suspended gold nano particles are commercially available in a variety of sizes, shapes, and chemical modifications for ease of conjugation with biomolecular recognition molecule. This property is applied to biosensors by coating the nano particles with an antibody and measuring the spectrum shift when the antibody captures the antigen.

Nanosphere lithography for LSPR devices Gold nano particles can also be fabricated on a substrate by physical methods. For low-cost and large-area periodic patterns of nano particles, nanosphere lithography (NSL) is used (Yonzon et al., 2004). In NSL, a commercial polystyrene latex nanosphere is deposited on a substrate to formed a closely packed-two-dimensional hexagonally array. Then, gold metal is vapor deposited in vacuum. The nano-sphere array works as a shadow mask in the vapor deposition. Because the metal vapor goes straight in a vacuum, the metal vapor reaches the substrate through the gaps between packed spheres. After nano-sphere mask is removed, the metal pattern shows gold nano-triangles of a hexagonal array. In that work, when nano-spheres with diameters of 390 nm were used, nanoparticles with a width of  100 nm were fabricated. A comparison of the biosensor performance between an LSPR device fabricated by NSL and an SPR device showed that RI sensitive length was much shorter in LSPR (5–6 nm) than in SPR (200 nm). They suggested that molecules giving an RI far from that gold surface in SPR may not be specifically binding molecules (Yonzon et al., 2004).

Physical Sensors: Plasmonic Sensors

55

Fig. 4 (AD) TEM images, (E) corresponding vis-NIR spectra, and (F) photographs of the glutathione-mediated gold nano rods prepared at different concentrations of glutathione: 0, 75, 150, and 350 nM, respectively. Modified from with permission He M-Q, Chen S, Yao K, et al. (2020). Precisely tuning lspr property via “peptide-encoded” morphological evolution of gold nanorods for quantitative visualization of enzyme activity. Analytical Chemistry 92: 1395–1401. Copyright (2020) American chemical Society.

Bottom-up fabrication for LSPR devices By using electron beam lithography, arbitrary two-dimensional shapes and arrangements of nano particles can be fabricated and optimized for specific applications. A nano particle can be considered a kind of antenna that collects and transmits an electromagnetic wave in specific conditions. The spectrum of gold nano particles is calculated from finite differential time domain (FDTD) analysis (Agharazy Dormeny et al., 2020). From the FDTD analysis, the performance of the LSPR highly depends on the particle size and shape. Especially, a bowtie shape is expected to be a sensitive and controllable structure for biosensors, but the fabrication of such nano scale gold patterns is technically difficult (Kinkhabwala et al., 2009). In 2013, a nano scale-antenna array with the center gap as narrow as the 5 nm was reported (Dodson et al., 2013). Compared to SPR, LSPR is actively researched more fundamentally in theoretical designs with fabrication techniques (Ge et al., 2020), and also a commercial instrument is developed.

Imaging SPR In the Kretschmann configuration, we can use an collimated beam to obtain images mapped by RI on the gold film (Zeng et al., 2017). Fig. 2D shows this optical configuration. In the imaging SPR (called SPRi; 2D-SPR), the reflection intensity is used to measuring the RI. As shown in Fig. 3B, the intensity of the reflected light (DRp) changes with a small change in the layer 3 RI at a specific incident angles lower than resonance incident angle. Therefore, we can measure the n3 from reflected light intensity without searching for resonance position. This is called intensity interrogation mode. When the sample surface is illuminated with a collimated light beam at the specific incident angle, the reflected light intensity profile directly correlates the refractive index distribution on layer 3. The RI sensitivity is determined by the slope of the SPR curve at the fixed incident angle. The dynamic range of the measurable RI is limited within the negative slope region in Fig. 3B. Because the sample is on the gold film and the optical path is on the other side of the glass, this microscope is useful for observing living cells (Horii et al., 2011) or operating microfluidics (Iwasaki et al., 2006). This configuration can provide focused image in a narrow area perpendicular to the beam direction, but the image is distorted in the beam direction because the reflection surface is observed from a very oblique angle. To improve the image quality and spatial resolution, SPR imaging with an objective lens was developed (Kano and Knoll, 1998; Meyer et al., 2012; Watanabe et al., 2012). Fig. 5 shows the optical configurations in the setup objective lens SPR imaging. Instead of

56

Physical Sensors: Plasmonic Sensors

Fig. 5 Schematic drawings of the combined surface plasmon microscope (SPM) setup for LISPM (lens-imaging-type SPM). The insets show the detailed optical path of the light nearby the sample. The microscope is switchable by flipping mirrors (FM1 and FM2). BE, beam expander; FM, flipping mirror; L, lens; BS, beam splitter; M, mirror; OBJ, microscope objective; BFP, back focal plane of OBJ. Modified from with permission Toma K, Kano H and Offenhäusser A (2014) Label-free measurement of cell–electrode cleft gap distance with high spatial resolution surface plasmon microscopy. ACS Nano 8: 12612–12619. Copyright (2014) American chemical Society.

a prism, an oil immersion high numerical aperture (NA) (> 1.4) objective lens is used to illuminate the gold surface. The polarized incident light passes through the peripheral of the objective lens and is reflected at near the lens axis on the gold film. At this point the incident light is a parallel beam. The reflected beam is collected by the same objective lens, passed through the focusing lens and imaged on to the CCD plane. This setup provides a correct aspect ratio and forced image with spatial resolution of a normal microscope. The image obtained with the setups in Figs. 2D and 5 is an RI-contrast image on the gold. Since the RI mapping does not need a labeling step in advance, imaging SPR microscopy is suitable for real-time whole-cell imaging (Toma et al., 2014). This SPR imaging configuration has been used to monitor DNA and RNA hybridization in real time (Halpern et al., 2014; Aoki et al., 2019). In addition to SPR imaging with the Kretschmann configuration, macroscale LSPR imaging has been also achieved (Raphael et al., 2013; Ruemmele et al., 2013). In this technique, islands comprising arrays of nano-particles are fabricated on a glass slide. Each island consists of gold nano disks. The size of islands is suitable for immobilization of molecular recognition molecules by mechanical spotting. The light transmitted through the LSPR glass plate is then filtered by a wavelength selector, and the plate image is focused on a CCD. The synchronized operation of the filter and CCD yields the spectrum of individual LSPR particle islands. Ref. Zeng et al. (2019) reported mechanical angle scanning SPR imaging. This method constructs an image by raster scanning the point reflection optics. By eliminating laser speckle, they showed RI sensitivity of 1.52e-6 and RI dynamic range of 4.6e-2. SPR imaging can image the RI change of biological binding events in a multiplexing manner.

One-dimensional SPR The Kretschmann SPR configuration in Fig. 2A can be parallelized by extending the optics perpendicular to the paper plane. A cylindrical prism and a wedge shaped beam that is line-focused on the reflection plane are used for RI sensing. This configuration can measure one-dimensional (1D) RI distributions in the angle interrogation mode (Horiuchi et al., 2012a,b). Therefore, high dynamic range and sensitivity can stand side by side with 1D multiplicity. Ref. (Sugai et al., 2020) used the real-time and multiplex measurement capability of 1D SPR system for feedback control of cell culture.

Fiber/waveguide based SPR Optical fiber and optical waveguide have core cladding structure to transmit light. They were originally developed for telecommunications technology, and the knowledge regarding their design and fabrication is vast. Moreover, fiber and waveguide are components are commercially available. Therefore, optical fiber has been exploited for SPR since the very beginning (Homola et al., 1999; Sharma et al., 2018). Fiber-based and waveguide SPR sensor can be realized by coating gold thin film or LSPR particle on the exposed core surface at the end of fiber or the middle of the fiber. When polychromatic light is used, an SPR curve is observed in the transmission spectrum. When a suitable monochromatic light is transmitted, the RI of sample can be given from transmittance. Fiber-based SPR can use the same chemistry as prism-based ones, and the sensing part can be made smaller than in prismbased SPR. A fiber SPR system can separate the light source and detection module from the sensing part and position them away from it. This is useful for washing the sensing part. Even the sensing part can be deformed and bent. Such characteristics of these devices have been demonstrated in detection of toxic targets (Xu et al., 2018).

Physical Sensors: Plasmonic Sensors

57

Portable SPR The optical configuration for SPR is rather simple. This is especially true for the Kretschmann configuration, which can be constructed with solid state electronic devices and fixed optical elements. Therefore, implementations of SPR in portable sensors has been anticipated (Chinowsky et al., 2007). Ref. Piliarik and Homola (2009) reported a compact SPR sensor operating in the wavelength interrogation mode. They employed gratings that simultaneously couple incident light into surface plasmons and disperses them in a spectrum. The system could measure RI change of 3e-7. With recent rapid advancements in the high-resolution imaging and powerful information processing capabilities of smartphones and low-cost, fast-turnaround prototyping by 3D printers, many types of portable SPR sensors have been proposed. In a recent review, the capability of SPR for portable devices is discussed comprehensively (Masson, 2020). Ref. Guner et al. (2017) reported an SPR imaging system for on-site immunoassay. The light source and optics were integrated with 3D-printed housing, and a smartphone’s imaging capability is used as detection optics and for signal processing. Ref. (Walter et al., 2020) reported an all-polymer planar waveguide SPR chip. The light source and imaging camera of a smartphone were used as the optical element of RI sensor base on the wavelength interrogation mode. The planar waveguide was fabricated by embossing a common optical plastic (PMMA, polymethyl methacrylate) for the cladding structure and filling the core part with UV curable epoxy. The optical coupling of the waveguide and the smartphone was accomplished by using a tapered mirror and grating that were integrated in the waveguide.

Selectivity and sensitivity of SPR sensors Selectivity SPR and LSPR devices can detect small levels of RI change resulting from the molecular binding event on a thin layer, but they are not selective in how the RI change is caused. The RI may change by many environmental conditions, such as temperature and solution concentration. Moreover, the refractive index of the SPR device surface can be altered by molecular adsorption and detachment by the undesired interaction of molecules with the device material. For example, on a gold surface, amino groups containing sulfur atom form a week valence bond, and the negatively charged ions adsorbed the surface. Such adsorbing molecules will cause a relatively large RI change in the SP device surface and hinder the small change in biosensor functionality. Therefore, surface inactivation for unwanted molecules (nonspecifically adsorbing molecules) is required for biosensor applications (Takasu et al., 2015). By using multichannel SPR or imaging SPR, the environmentally induced changes in RI can be compensated by comparing the RI of channels of the active and non-active part of the surface to the detection target molecule ( Springer et al., 2013). This method is widely used in the Kretschmann configuration, because of its wide RI dynamic range.

Combination of SPR and Raman spectroscopy In an emerging utilization of SPR, an enhanced electric field is used for fluorescence (Kinkhabwala et al., 2009), and Raman scattering or surface enhanced Raman scattering (SERS) spectroscopy (Futamata, 1997; Sawai et al., 2007; Meyer et al., 2012; Campu et al., 2018; Pin et al., 2018; Cai et al., 2021; Zhang et al., 2020) are reported. Because Raman spectroscopy can identify the chemical composition of the target by vibration spectroscopy, and SPR with molecular recognition material gives highly specific determination of complex target molecules, their combination complements each other to provide rapid and quantitative detection. Moreover, not specifically binding molecules on the SPR sensing surface always complicate the biological measurement. To eliminate this unwanted non-specific binding, a washing step and blocking reagents are used in many cases. Raman spectroscopy can help determining the ratio of nonspecific binding. The optical setups for the SERS/SPR combination can be a single Kretschmann configuration, using a prism or immersing a highNA objective microscopy lens and a Weierstrass (hyper-hemisphere) prism to collect scattered light. This optical setup has been used for sequence selective hybridization of single-stranded DNA (Halpern et al., 2014). For other optical techniques combinations with SPR, when the sample layer molecules show chirality, the p-polarized incident light reflected at the chiral layer will have s-polarized light, and the chirality of thin film can be measured (Mi and Van, 2014; Lu et al., 2017; Droulias and Bougas, 2019).

Sensing volume, sensitivity of SPR As show in Fig. 3B, the SPR curve shifts with the RI of the sample layer n3. The critical angle does not change with n3 (RI of 50 nm in sample layer), it is determined mainly by n4 (RI of bulk water). This localized sensitivity is a prominent characteristic of SPR measurement. RI measurement by SPR is a type of peak tracking detection. Therefore, the incident light intensity and the received light intensity itself are basically not relevant to the detection sensitivity (except for fixed incident angle intensity interrogation mode); n3 is determined from the shape and position of the SPR curve. The light and plasmon interaction is confined in a small volume, typically 2e-17 m 3. Therefore, the required path length for the probe light is significantly reduced compared to light adsorption measurement or the non-resonant attenuation total reflection (ATR) spectroscopy, for which multiple reflections are

58

Physical Sensors: Plasmonic Sensors

needed to increase the optical path length. In the case of SPR measurement, the probe (incident) light is reflected once at the SP medium, and produces enough extinction of light for determining the resonance points (incident angle or wavelength). The RI resolution or sensitivity of SPR measurement has been repeatedly compared and discussed in review articles. The common understanding of SPR RI sensitivity is that may fall in 1e-5-1e-7. Although the SPR angle can be converted to the absolute RI using Fresnel reflection theory and calculation, how the parameters affect the absolute RI value is uncertain at the resolution of 1e-7. For example, the RI of gold film will vary from the preparation batch and with thickness, roughness and uniformity of the film. These are practically difficult to control to such precision. Therefore, high resolution of the RI value derived from SPR is always a relative change (differential) in the RI of the gold surface over certain time period. The quantitative SPR curve shift by caused by the n3 in Fig. 2A change is determined by the optical properties of gold film. It is not relevant to the precision of optical configurations. Therefore, instrumental stabilization and noise reduction in electronic devices are essential to the higher RI sensitivity in SPR measurement (Kolomenskii et al., 1997). For example, Ref. Zybin et al. (2005) described two-wavelength incident beam technique SPR for the sensitivity improvement. They used two laser diodes (785 and 830 nm) and multiplexing with a beam splitter, and switched them at 2 kHz and illuminated the gold film. The reflected light was detected by a photodiode array synchronized with the wavelength switching. The calibration curve for sucrose concentration was linear up to a 2e-3 RI increase and the RI resolution of 5e-6 for sucrose solution.

Noise source Because the RI resolution of SPR is very high, the time fluctuation of SPR data is significant when the sample is unstable. Even if the sample RI is stable, SPR data show time-dependent noise. The source of such noise has been investigated systematically (Chiang et al., 2007). The RI of water depends on its temperature rather significantly (Luo et al., 2019). Moreover, the coefficient of RI vs. temperature differs considerably from molecule to molecule including the sign (Bashkatov and Genina, 2003; Özdemir and Turhan-Sayan, 2003). Device temperature stabilization and the sample temperature become critical as measurement sensitivity/ resolution increases (low-concentration measurement). Ref. (Zybin et al., 2005) showed the dependence of the actual SPR signal shift on sample temperatures from 19 to 60  C. This temperature compensation capability was used for detection of double strand DNA unfolding by heating at 90  C. In a special condition, using a stabilized laser and photodiodes for the signal and reference and frequency-domain filter, detection of as low as a 1e-9 RI (1 fg/mm2) of air in response to pressure change was demonstrated with a Kretschmann configuration (Davis and Wilson, 2000; Wang et al., 2011). This was reported shot noise limit regime.

Mass transport In detecting low-concentration molecules in a small sample volume, the addressability of the target molecule to the sensing volume becomes critical. Different from the optical system of the plasmon sensor device itself, the mass transport of target molecules to the sensing surface effectively limits the detection at the low-concentration end. In biosensor applications of SPR, the sample is commonly fed through a microfluidic device. This is because a biological sample is in aqueous solution, the sample volume is limited, and for the molecular interaction study, the sensor surface should be regenerated by detaching and washing with several reagents, and continuous automatic operation is required. In the laminar flow condition, which is common for microfluidic channels, the flow velocity distribution is parabolic with the maximum velocity at the center of the channel cross-section and zero at the channel wall. Because the SPR has sensitivity very near (< 200 nm) the plasmon medium surface (gold thin film or nano particles), the flow velocity is almost zero for the SPR sensing surface and the diffusional transport becomes rate determining even in a sample is in the flow condition. When the binding rate of sample molecules is high enough compared the diffusional transport, the concentration of sample molecules on the SPR sensing surface becomes effectively lower than the fed concentration. This causes errors in determining the kinetic information and sample concentration (Myszka et al., 1998), and the experimental conditions (flow rate, sensing surface activity) should be carefully designed. This effect is more pronounced in LSPR because the sensing region is more compact than in SPR. Ref. ( Spacková et al., 2018) studied the effect of changing the island density of gold rods. By determining the lowest concentration detectable by adsorption rate measurement of DNA, lower island density was found to give the lower detectable concentrationdeven the plasmon resonance signal became shallower as the island density decreased. This showed that the reduction of sensing surface area will also reduce the effect stemming from the diffusion-limited condition. Therefore, improving the sensor performance requires device optimization including device design (sensing surface area), the sample liquid flow condition (flow rate, flow cell size), immobilization density, light source, and readout electronics (CCD).

Data processing in SPR SPR angle/wavelength calculation from SPR curve is a critical step in SPR data processing. An observed SPR curve will follow the Fresnel multilayer reflection formula, so it is natural to fit the experimentally obtained SPR curve to the Fresnel formula. Because the number of layers and their thickness and complex RI are not known, the fitting is ambiguous, and it is hard to obtain reasonable results. The second choice maybe approximating to the SPR curve by a quadratic polynomial and derive the SPR angle. This gives

Physical Sensors: Plasmonic Sensors

59

a deterministic result; however, because this approximation is poor, the calculation result does not show a monotonic relation to the continuous change of RI. The other method is to calculate the centroid of the SPR curve dip (Johansen et al., 2000). This calculation is mere averaging to SPR curve, so it very simple and fast. Moreover, the noise level usually lower than in the quadratic approach (Piliarik and Homola, 2009;  Spacková et al., 2018). Another model-based approach for approximating the SPR curve by a five-parameter rational function has been proposed (Kurihara et al., 2002). This approach needs nonlinear curve fitting, so the calculation results depend on the initial guess and residual cut off value. Ref. Etxebarria-Elezgarai et al. (2020) reconsidered the Kretschmann configuration and proposed beam shaping and data processing for better RI resolution. When the chemical analysis aims for a qualitative judgment and a quantitative RI change is not necessary, the required result can be calculated from the raw SPR curve or spectrum from the SPR device. In such a case, a machine learning algorithm can be used. This approach has emerged recently (Ballard et al., 2017; Moon et al., 2019). This method also helps simplify the measurement protocol through the use of low-cost optical devices. The computational capability of small computers and smartphones increasing at a rapid pace, so the use of machine learning is advantageous in portable sensor systems.

Comparison to other methods Ref. Fan et al. (2008) compared SPR with other methods. Compared to optical devices such as ring resonators and waveguide-based interferometers for detecting changes in RIs, SPR devices are the simplest in term of device structure. The RI change can be considered as a mass change. Compared to QCM (quartz crystal microbalance) (Höök et al., 2001) and MEMS (Micro electro mechanical systems) cantilevers, SPR devices enable the easiest multiplexing of the measurements and are less susceptive to sample properties. For biological sensing applications, SPR can be used in aqueous solution and even in biological fluids without severe limitations (viscosity, solid mixture) in the measurement.

Future perspective The development of the SPR devices have been boosted by innovations in electro-optical devices (CCDs, CMOS camera, LDs, LEDs) and microfluidics. But, the absolute sensitivity of SPR for RIs has little improved since the 1990s. This is because the principle of SPR is solely based on the Fresnel reflection theorem and optical property of gold. This in turn gives a stable and solid sensing principle. In biosensing, further improvement of sensitivity will soon face the discrimination of nonspecific binding. Recent developments in the combination of SPR and other spectroscopic methods are expected to overcome such problems. In applications on-site measurements and point of care testing, SPR devices will continue to be good candidates as transducers of internet of things devices.

References Agharazy Dormeny, A., Abedini Sohi, P., Kahrizi, M., 2020. Design and simulation of a refractive index sensor based on SPR and LSPR using gold nanostructures. Results in Physics 16, 102869. Aoki, H., Corn, R.M., Matthews, B., 2019. MicroRNA detection on microsensor arrays by spr imaging measurements with enzymatic signal enhancement. Biosensors and Bioelectronics 142, 111565. Ballard, Z.S., Shir, D., Bhardwaj, A., et al., 2017. Computational sensing using low-cost and mobile plasmonic readers designed by machine learning. ACS Nano 11, 2266–2274. Bashkatov, A., Genina, E., 2003. Water refractive index in dependence on temperature and wavelength: A simple approximation. In: Saratov Fall Meeting 2002 Laser Physics and Photonics, Spectroscopy, and Molecular Modeling III; Coherent Optics of Ordered and Random Media IIISPIE. Bosch, I., De Puig, H., Hiley, M., et al., 2017. Rapid antigen tests for dengue virus serotypes and zika virus in patient serum. Science Translational Medicine 9, eaan1589. Cai, Z., Hu, Y., Sun, Y., et al., 2021. Plasmonic sers biosensor based on multibranched gold nanoparticles embedded in polydimethylsiloxane for quantification of hematin in human erythrocytes. Analytical Chemistry 93 (2), 1025–1032. Campu, A., Lerouge, F., Chateau, D., et al., 2018. Gold nanobipyramids performing as highly sensitive dual-modal optical immunosensors. Analytical Chemistry 90, 8567–8575. Chiang, H.P., Chen, C.W., Wu, J.J., et al., 2007. Effects of temperature on the surface plasmon resonance at a metal-semiconductor interface. Thin Solid Films 515, 6953–6961. Chinowsky, T.M., Quinn, J.G., Bartholomew, D.U., et al., 2003. Performance of the spreeta 2000 integrated surface plasmon resonance affinity sensor. Sensors and Actuators B: Chemical 91, 266–274. Chinowsky, T.M., Grow, M.S., Johnston, K.S., et al., 2007. Compact, high performance surface plasmon resonance imaging system. Biosensors and Bioelectronics 22, 2208–2215. Choi, I., Choi, Y., 2011. Plasmonic nanosensors: Review and prospect. IEEE Journal of Selected Topics in Quantum Electronics 18, 1110–1121. Davis, T.M., Wilson, W.D., 2000. Determination of the refractive index increments of small molecules for correction of surface plasmon resonance data. Analytical Biochemistry 284, 348–353. Dodson, S., Haggui, M., Bachelot, R., et al., 2013. Optimizing electromagnetic hotspots in plasmonic bowtie nanoantennae. The Journal of Physical Chemistry Letters 4, 496–501. Droulias, S., Bougas, L., 2019. Surface plasmon platform for angle-resolved chiral sensing. ACS Photonics 6, 1485–1492. Etxebarria-Elezgarai, J., Mowat, M., Lopez, E., et al., 2020. Gaussian beam shaping and multivariate analysis in plasmonic sensing. Analytical Chemistry 92 (24), 16236–16244. Fan, X., White, I.M., Shopova, S.I., et al., 2008. Sensitive optical biosensors for unlabeled targets: A review. Analytica Chimica Acta 620, 8–26. Fukuda, N., Tithimanan, S., Ushijima, H., Yamamoto, N., 2017. Paper based plasmonic surface for chemical biosensing by the attenuated total reflection method. MRS Advances 2, 2303–2308. Futamata, M., 1997. Application of attenuated total reflection surface-plasmon-polariton Raman spectroscopy to gold and copper. Applied Optics 36, 364–375. Gawedzinski, J., Pawlowski, M., Tkaczyk, T., 2017. Quantitative evaluation of performance of three-dimensional printed lenses. Optical Engineering 56, 084110. Ge, D., Marguet, S., Issa, A., et al., 2020. Hybrid plasmonic nano-emitters with controlled single quantum emitter positioning on the local excitation field. Nature Communications 11, 3414.

60

Physical Sensors: Plasmonic Sensors

Ghosh, S.K., Nath, S., Kundu, S., et al., 2004. Solvent and ligand effects on the localized surface plasmon resonance (lspr) of gold colloids. The Journal of Physical Chemistry B 108, 13963–13971. Guner, H., Ozgur, E., Kokturk, G., et al., 2017. A smartphone based surface plasmon resonance imaging (SPRi) platform for on-site biodetection. Sensors and Actuators B: Chemical 239, 571–577. Halpern, A.R., Wood, J.B., Wang, Y., Corn, R.M., 2014. Single-nanoparticle near-infrared surface plasmon resonance microscopy for real-time measurements of DNA hybridization adsorption. ACS Nano 8, 1022–1030. He, M.-Q., Chen, S., Yao, K., et al., 2020. Precisely tuning lspr property via “peptide-encoded” morphological evolution of gold nanorods for quantitative visualization of enzyme activity. Analytical Chemistry 92, 1395–1401. Hinman, S.S., McKeating, K.S., Cheng, Q., 2017. Plasmonic sensing with 3d printed optics. Analytical Chemistry 89, 12626–12630. Homola, J., 2003. Present and future of surface plasmon resonance biosensors. Analytical and Bioanalytical Chemistry 377, 528–539. Homola, J., Yee, S.S., Gauglitz, G., 1999. Surface plasmon resonance sensors: Review. Sensors and Actuators B: Chemical 54, 3–15. Höök, F., Kasemo, B., Nylander, T., et al., 2001. Variations in coupled water, viscoelastic properties, and film thickness of a Mefp-1 protein film during adsorption and cross-linking: A quartz crystal microbalance with dissipation monitoring, ellipsometry, and surface plasmon resonance study. Analytical Chemistry 73, 5796–5804. Horii, M., Shinohara, H., Iribe, Y., Suzuki, M., 2011. Living cell-based allergen sensing using a high resolution two-dimensional surface plasmon resonance imager. Analyst 136, 2706–2711. Horiuchi, T., Miura, T., Iwasaki, Y., et al., 2012a. Passive fluidic chip composed of integrated vertical capillary tubes developed for on-site spr immunoassay analysis targeting real samples. Sensors 12, 7095–7108. Horiuchi, T., Tobita, T., Miura, T., et al., 2012b. Floating chip mounting system driven by repulsive force of permanent magnets for multiple on-site spr immunoassay measurements. Sensors 12, 13964–13984. Inoue, S., Seyama, M., Miura, T., et al., 2016. A reliable aptamer array prepared by repeating inkjet-spotting toward on-site measurement. Biosensors and Bioelectronics 85, 943–949. Iwasaki, Y., Horiuchi, T., Morita, M., Niwa, O., 1997. Time differential surface plasmon resonance measurements applied for electrochemical analysis. Electroanalysis 9, 1239–1241. Iwasaki, Y., Niwa, O., Morita, M., 1999. Electrochemical reaction of cytochrome c on polyaspartic acid modified gold electrodes. Sensors and Materials 11, 51–56. Iwasaki, Y., Horiuchi, T., Niwa, O., 2001. Detection of electrochemical enzymatic reactions by surface plasmon resonance measurement. Analytical Chemistry 73, 1595–1598. Iwasaki, Y., Tobita, T., Kurihara, K., et al., 2006. Imaging of flow pattern in micro flow channel using surface plasmon resonance. Measurement Science and Technology 17, 3184–3188. Johansen, K., Stålberg, R., Lundström, I., Liedberg, B., 2000. Surface plasmon resonance: Instrumental resolution using photo diode arrays. Measurement Science and Technology 11, 1630–1638. Kano, H., Knoll, W., 1998. Locally excited surface-plasmon-polaritons for thickness measurement of LBK films. Optics Communications 153, 235–239. Kessler, M.A., Hall, E.a.H., 1996. Multilayered structures exhibiting long-range surface exciton resonance. Thin Solid Films 272, 161–169. Kinkhabwala, A., Yu, Z., Fan, S., et al., 2009. Large single-molecule fluorescence enhancements produced by a bowtie nanoantenna. Nature Photonics 3, 654–657. Kolomenskii, A.A., Gershon, P., Schuessler, H., 1997. Sensitivity and detection limit of concentration and adsorption measurements by laser-induced surface-plasmon resonance. Applied Optics 36, 6539–6547. Kravets, V.G., Kabashin, A.V., Barnes, W.L., Grigorenko, A.N., 2018. Plasmonic surface lattice resonances: A review of properties and applications. Chemical Reviews 118, 5912–5951. Kurihara, K., Nakamura, K., Suzuki, K., 2002. Asymmetric SPR sensor response curve-fitting equation for the accurate determination of SPR resonance angle. Sensors and Actuators B: Chemical 86, 49–57. Lambert, A., Valiulis, S., Cheng, Q., 2018. Advances in optical sensing and bioanalysis enabled by 3d printing. ACS Sensors 3, 2475–2491. Lertvachirapaiboon, C., Baba, A., Shinbo, K., Kato, K., 2021. Dual-mode surface plasmon resonance sensor chip using a grating 3D-printed prism. Analytica Chimica Acta 1147, 23–29. Lu, J., Chang, Y.-X., Zhang, N.-N., et al., 2017. Chiral plasmonic nanochains via the self-assembly of gold nanorods and helical glutathione oligomers facilitated by cetyltrimethylammonium bromide micelles. ACS Nano 11, 3463–3475. Lucarini, V., Saarinen, J.J., Peiponen, K.-E., Vartiainen, E.M., 2005. Kramers-Kronig Relations in Optical Materials Research. Springer Science & Business Media. Luo, W., Wang, R., Li, H., et al., 2019. Simultaneous measurement of refractive index and temperature for prism-based surface plasmon resonance sensors. Optics Express 27, 576–589. Maier, S.A., Brongersma, M.L., Kik, P.G., et al., 2001. PlasmonicsdA route to nanoscale optical devices. Advanced Materials 13, 1501–1505. Masson, J.-F., 2017. Surface plasmon resonance clinical biosensors for medical diagnostics. ACS Sensors 2, 16–30. Masson, J.-F., 2020. Portable and field-deployed surface plasmon resonance and plasmonic sensors. Analyst 145, 3776–3800. Meyer, S.A., Auguié, B., Le Ru, E.C., Etchegoin, P.G., 2012. Combined SPR and SERS microscopy in the Kretschmann configuration. The Journal of Physical Chemistry A 116, 1000–1007. Mi, G., Van, V., 2014. Characteristics of surface plasmon polaritons at a chiral–metal interface. Optics Letters 39, 2028–2031. Moon, G., Son, T., Lee, H., Kim, D., 2019. Deep learning approach for enhanced detection of surface plasmon scattering. Analytical Chemistry 91, 9538–9545. Myszka, D.G., He, X., Dembo, M., et al., 1998. Extending the range of rate constants available from biacore: Interpreting mass transport-influenced binding data. Biophysical Journal 75, 583–594. Nootchanat, S., Jaikeandee, W., Yaiwong, P., et al., 2019. Fabrication of miniature surface plasmon resonance sensor chips by using confined sessile drop technique. ACS Applied Materials & Interfaces 11, 11954–11960. Özdemir, S.A.K., Turhan-Sayan, G., 2003. Temperature effects on surface plasmon resonance: Design considerations for an optical temperature sensor. Journal of Lightwave Technology 21, 805. Pan, M.-Y., Lee, K.-L., Lo, S.-C., Wei, P.-K., 2018. Resonant position tracking method for smartphone-based surface plasmon sensor. Analytica Chimica Acta 1032, 99–106. Piliarik, M., Homola, J., 2009. Surface plasmon resonance (spr) sensors: Approaching their limits? Optics Express 17, 16505–16517. Pin, C., Ishida, S., Takahashi, G., et al., 2018. Trapping and deposition of dye–molecule nanoparticles in the nanogap of a plasmonic antenna. ACS Omega 3, 4878–4883. Raphael, M.P., Christodoulides, J.A., Delehanty, J.B., et al., 2013. Quantitative LSPR imaging for biosensing with single nanostructure resolution. Biophysical Journal 104, 30–36. Ruemmele, J.A., Hall, W.P., Ruvuna, L.K., Van Duyne, R.P., 2013. A localized surface plasmon resonance imaging instrument for multiplexed biosensing. Analytical Chemistry 85, 4560–4566. Sawai, Y., Takimoto, B., Nabika, H., et al., 2007. Observation of a small number of molecules at a metal nanogap arrayed on a solid surface using surface-enhanced Raman scattering. Journal of the American Chemical Society 129, 1658–1662. Schasfoort, R.B., 2017. Handbook of Surface Plasmon Resonance. Royal Society of Chemistry. Sharma, A.K., Pandey, A.K., Kaur, B., 2018. A review of advancements (2007–2017) in plasmonics-based optical fiber sensors. Optical Fiber Technology 43, 20–34. Shrestha, B., Pipatpanukul, C., Houngkamhang, N., et al., 2020. Application of printable antibody ink for solid-phase immobilization of abo antibody using photoactive hydrogel for surface plasmon resonance imaging. Sensors and Actuators B: Chemical 320, 128358. Spacková, B., Lynn, N.S., Slabý, J., et al., 2018. A route to superior performance of a nanoplasmonic biosensor: Consideration of both photonic and mass transport aspects. ACS Photonics 5, 1019–1025.

Physical Sensors: Plasmonic Sensors

61

Springer, T., Bocková, M., Homola, J., 2013. Label-free biosensing in complex media: A referencing approach. Analytical Chemistry 85, 5637–5640. Sugai, H., Tomita, S., Ishihara, S., et al., 2020. Microfluidic sensing system with a multichannel surface plasmon resonance chip: Damage-free characterization of cells by pattern recognition. Analytical Chemistry 92, 14939–14946. Takasu, K., Kushiro, K., Hayashi, K., et al., 2015. Polymer brush biointerfaces for highly sensitive biosensors that preserve the structure and function of immobilized proteins. Sensors and Actuators B: Chemical 216, 428–433. Tanabe, I., Tanaka, Y.Y., Watari, K., et al., 2017. Aluminum film thickness dependence of surface plasmon resonance in the far-and deep-ultraviolet regions. Chemistry Letters 46, 1560–1563. Toma, K., Kano, H., Offenhäusser, A., 2014. Label-free measurement of cell–electrode cleft gap distance with high spatial resolution surface plasmon microscopy. ACS Nano 8, 12612–12619. Vigderman, L., Zubarev, E.R., 2013. High-yield synthesis of gold nanorods with longitudinal SPR peak greater than 1200 nm using hydroquinone as a reducing agent. Chemistry of Materials 25, 1450–1457. Walter, J.-G., Alwis, L.S., Roth, B., Bremer, K., 2020. All-optical planar polymer waveguide-based biosensor chip designed for smartphone-assisted detection of vitamin d. Sensors 20, 6771. Wang, X., Jefferson, M., Hobbs, P.C., et al., 2011. Shot-noise limited detection for surface plasmon sensing. Optics Express 19, 107–117. Watanabe, K., Matsuura, K., Kawata, F., et al., 2012. Scanning and non-scanning surface plasmon microscopy to observe cell adhesion sites. Biomedical Optics Express 3, 354–359. Xu, H., Käll, M., 2002. Modeling the optical response of nanoparticle-based surface plasmon resonance sensors. Sensors and Actuators B: Chemical 87, 244–249. Xu, Y., Luo, Z., Chen, J., et al., 2018. U-shaped fiber-optic probe-based localized surface plasmon resonance biosensor for real-time detection of salmonella typhimurium. Analytical Chemistry 90, 13640–13646. Yonzon, C.R., Jeoung, E., Zou, S., et al., 2004. A comparative analysis of localized and propagating surface plasmon resonance sensors: The binding of concanavalin a to a monosaccharide functionalized self-assembled monolayer. Journal of the American Chemical Society 126, 12669–12676. Zeng, Y., Hu, R., Wang, L., et al., 2017. Recent advances in surface plasmon resonance imaging: Detection speed, sensitivity, and portability. Nanophotonics 6, 1017–1030. Zeng, Y., Zhou, J., Xiao, X., et al., 2019. A speckle-free angular interrogation spr imaging sensor based on galvanometer scan and laser excitation. Plasmonics 14, 1497–1504. Zhang, D., You, H., Zhang, L., Fang, J., 2020. Facile surface modification of mesoporous Au nanoparticles for highly sensitive SERS detection. Analytical Chemistry 92, 15379– 15387. Zybin, A., Grunwald, C., Mirsky, V.M., et al., 2005. Double-wavelength technique for surface plasmon resonance measurements: Basic concept and applications for single sensors and two-dimensional sensor arrays. Analytical Chemistry 77, 2393–2399.

Physical Sensors: Mechanical Sensors Satoshi Saga, Kumamoto University, Kumamoto, Japan © 2023 Elsevier Ltd. All rights reserved.

Introduction Examples of conventional mechanical sensors Force sensors Sensors as mathematical function Injection (one-to-one function) Wide domain and range Effects of other parameters Measurement independency Independency of the target value history Temperature independence What kind of mechanical sensors are needed for measuring contact state? Contact sensors of human Bidirectionality Spatial distribution Range, dimension, and material Some devices for force distribution Frequency distribution Information transfer problem Imaging devices for information transfer Other novel devices for information transfer Is every distribution required? Deep learning for haptic information Classification of haptic information using deep learning Acquisition of tactile information and their classification using machine learning Classification as a function of sensors Generation of sensing information by deep learning Conclusion References

62 63 63 64 65 65 65 65 65 66 67 67 68 68 69 69 70 70 70 71 71 71 71 71 72 72 74 74

Abstract To design the use of the sensors, we have to consider how to measure the environment. In this article, we have introduced and reviewed principles, functionality, and future of mechanical sensors. First, we have considered what the mechanical sensors are. Mechanical sensors measure the displacement/movement/direction/acceleration of measured target. The measurement methods vary, though, we have picked up the sensors which measure the mechanical changes. First, we have reviewed the mechanical sensors from a conventional viewpoint, mathematical viewpoint, and haptic information viewpoint. Then, we discuss the potential of the machine learning as the future of mechanical sensors. The regression layer in the Convolutional Neural Network model can evaluate the information which is in the input data. By using the Generative Adversarial Network model, we can parameterize and sense the intuitive index with a more systematic method. For further multifunctional mechanical sensing, machine learning method will be essential technology.

Introduction Nowadays, there are many kinds of sensors, for example, position, angle, velocity, accelerometers, mass, force, temperature, humidity, pressure, intensity/color/frequency of light, voltage, current, resistance, capacitance, time, frequency, or so. These sensors detect events or changes in the target environment and quantify them by employing physical/chemical/biological changes. To evaluate the changes, many physical/chemical/biological phenomena are used. Further, most of these sensors quantify the changes in the environment as the electric changes. For example, change of light, temperature, pressure, velocity, acceleration, angle velocity, voltage, current, resistance, pH, or existence of chemical materials are converted to change of voltage, current, resistance, capacitance or inductance. To design the use of the sensors, we have to consider how to measure the environment. In this article, we introduce and review principles, functionality, and future of mechanical sensors. We consider what the mechanical sensors are. Mechanical sensors

62

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00031-5

Physical Sensors: Mechanical Sensors

63

measure the displacement/movement/direction/acceleration of measured target. Thus, the mechanical change of the target environment is important. The measurement methods vary, though, we pick up the sensors which measure the mechanical changes. First, we review the mechanical sensors from a conventional viewpoint, mathematical viewpoint, and haptic information viewpoint. Then we discuss the future of mechanical sensors.

Examples of conventional mechanical sensors To evaluate the changes of environment, many physical/chemical/biological phenomena are used. Nowadays, there are many kinds of sensors, for example, position, angle, velocity, accelerometers, mass, force, temperature, humidity, pressure, intensity/color/ frequency of light, voltage, current, resistance, capacitance, time, frequency, magnetic, gyroscope, etc. Further, most of these sensors quantify the changes in the environment as the electric changes. For example, change of light, temperature, pressure, velocity, acceleration, angle velocity, voltage, current, resistance, pH, or existence of chemical materials are converted to voltage, current, resistance, capacitance, or inductance. There are multiple measurement methods and ranges for measuring some physical parameters. For example, to measure the light intensity, we can employ the following sensors. Photodiodes, phototransistors, CCD/CMOS sensors, and photomultiplier tubes measure the light intensity using current/electric charges. CdS sensors measure the light intensity by the changes of resistance. As another example, if we want to measure the velocity of the fluid, we can employ small energy consumption by using a rotating body, Bernoulli’s principle by using a pitot tube, acoustic doppler current profiler, or the amount of losing heat to the fluid by using a hot-wire anemometer, etc. Here, we list the force/slippage/proximity sensing methods in Table 1; In this section, we overview several conventional mechanical sensors and their principles.

Force sensors Conventional contact sensors have been created from some principles of physics (Saga, 2008). They record the phenomena of the contact status using some physical principles. In order to record the deformation or stress information, many tactile sensors have been developed, e.g., strain gauge, piezoelectric effect, pressure-sensitive rubber, diaphragm, photometric pressure gauge, and SAW force sensor, etc.

Table 1

Force/slippage/proximity/thermal sensing methods.

Physical parameter

Principle

Examples

Force

strain gauge

strain gauge sensor (e.g., BL NanoSensor BL Autotec Ltd., 2020) pressure sensitive rubber (Shimojo et al., 2004), pressuresensitive elastomer

percolation phenomena contact resistance piezoelectric device capacitance surface acoustic wave light emission total reflection image sensing

Slippage Proximity

magnetic field ultrasound optical acceleration optical capacitance eddy current ultrasound

pressure-sensitive ink (e.g., FlexiForce Tekscan Inc, 2020), conductive fiber PVDF, PZT, crystal capacitive sensor (Schmitz et al., 2010; Medical Tactile Inc, 2020) SAW sensor (Benes et al., 1998) diffusive light (e.g., OptoForce Inc., 2020), optical fiber optical waveguide (Maekawa et al., 1993), transparent rubber (Saga et al., 2014) GelForce (Kamiyama et al., 2005), GelSight (Johnson and Adelson, 2009) hall effect (Jamone et al., 2015) acoustic resonance (Ando et al., 2001) displacement detection (Maldonado et al., 2012) vibration detection (Tremblay and Cutkosky, 1993) diffusive light (Terada et al., 2011), time of flight capacitive proximity sensor (Lee et al., 2009)

64



Physical Sensors: Mechanical Sensors Strain gauge: A load cell usually uses a strain gauge. Through a mechanical arrangement, the force being sensed deforms a strain gauge. The strain gauge converts the deformation (strain) to electrical signals. The electrical signal output is typically in the order of a few millivolts and requires amplification by an instrumentation amplifier before it can be used (Fig. 1).

A strain gauge takes advantage of the physical property of electrical conductance’s dependency. R¼

rl A

(1)

By differentiate the equation, we get

DR R

¼

Dr Dl DA r

þ



l

¼ ð1 þ 2sÞ

A

Dl

(2)

(3)

l

r: Resistance ratio, R: Resistance value, l: Length, A: Cross section, s: Poisson’s ratio. This ð1 þ2sÞ is called a gauge factor.



Piezoelectric device: A piezoelectric device uses a piezoelectricity effect. Piezoelectricity is the ability of some materials to generate an electric potential in response to applied mechanical stress. That is, this effect translates the strain information toward electric voltage. A PVDF also has piezoelectricity. By using this characteristic, some force sensors are created.

Dr r

¼ pE

Dl

(4)

l

With the Eqs. (2) and (4) we get;

DR R

¼ ðp þ 1 þ 2sÞ

Dl l

p : Piezoresistance coefficient; E : Young’s modulus

(5)

(6)

This ðp þ1 þ2sÞ is called a gauge factor.



Pressure-sensitive rubber: A pressure-sensitive rubber has been developed for the sheet-switch of the electronic circuits and has a unique property. It conducts electric current only when compressed and acts as an insulator when the pressure is released. This patented material is a composite of an elastomer and specially treated carbon particles and is available in gray-black flexible sheet form, 0.5 mm in thickness.

Sensors as mathematical function Sensor is like function in mathematics, which is a binary relation over two sets that associates every element of the first set (the domain of definition, domain), to exactly one element of the second set (range). The domain values are measured as raw data, and the range values are target values indicated by the measured raw data. The critical points of the ideal function are as follows.

Fig. 1

Strain gauge.

Physical Sensors: Mechanical Sensors

65

1. The function should be injective function, not multivalued function 2. The domain and the range is wide enough 3. Values of the domain are not affected by other parameters (a) Measurement independency (b) Independency of the target value history (c) Temperature independency

Injection (one-to-one function) The function of the sensor should be injective (Fig. 2). If the function is multivalued, one measured value indicates multiple values of the range. Thus the function cannot determine the actual value of the target. To employ such a multivalued function, we have to limit the range. There are some functions that indicate multivalued responses; however, we can use such a function by restricting the domain of definition and its range. We can get a single target value from the measured value in the limited range, which has a single relation between the domain and the range.

Wide domain and range The domain and the range of sensors should be wide enough. Here, quantization is the process of mapping input values from a large and/or continuous set to output values in a smaller and/or discrete set. Through quantization, measured values are rounded or truncated. In digital signal processing, the quantization processes are indispensable. Thus the quantization takes place in sensing. If the measurement domain is too small, quantization error becomes larger in the target range. When the minimum quantization unit is O Y, we cannot evaluate the range between O X (Fig. 3). On the other hand, the role of sensors is the indicator of the target. Thus a larger range is important. Even if the measurement domain is large, the smaller target range does not provide enough information. When the maximum domain is limited to

O Y, we only evaluate the range in O X.

Effects of other parameters Measurement independency Under the deviation method, sensors acquire change of the target by measurement. However, the measurement affect the sensing target, too. Thus, most of the sensors are designed small enough to limit the effect to the target (Yan, 2014). For example, alcohol thermometer measures the temperature of the target. To measure the target temperature, the thermometer uses the volume change of alcohol liquid, and the meniscus moves up the capillary of the thermometer. To change the volume, the temperature uses some quantity of heat. Thus, the heat quantity of the target is decreased by the thermometer. The decrease of the heat quantity induces the reduction of the target temperature. To reduce the effect of the dependency of the measurement, we have to make the heat capacity smaller (Fig. 4). On the other hand, zero method is a sensing method that adjusts until the measured quantity and the reference quantity become equal. In general, equilibrium can be detected with high accuracy, so that the measured quantity can be measured with the same accuracy as the reference quantity. In contrast to the deviation method in which energy is derived from the measurement target to move the needle of the instrument, energy is not deprived at the equilibrium point of the zero position method, and the effect on the measurement target is limited to a minimum. For example, the balance scale employs a zero method in mechanics. By adjusting an end of the balance with known weight, the balanced state shows the unknown weight on the other end equals the placed known weight. Wheatstone bridge is a zero method in electromagnetism that employs the balancing bridge of voltage. This method can measure precise resistance.

Independency of the target value history Hysteresis is the change in the state of a system depending not only on the current inputs but also on past inputs. Thus, hysteresis is the dependence of the state of a system on its history.

Fig. 2

Injective function.

66

Physical Sensors: Mechanical Sensors

Fig. 3 Relation between domain and range. Left: Small domain. Quantization error becomes larger in the target range, Right: small target; the smaller target range does not provide enough information.

Several hysteresis phenomena can be a dynamic lag between an input and an output that dis- appears if the input is varied more slowly; this is known as rate-dependent hysteresis. However, phenomena such as the magnetic hysteresis loops are mainly rateindependent, which makes a durable memory possible. The rate-dependent hysteresis mentioned above has a dependency on time. Such dependence is known as frequency response. Frequency response is the quantitative measure of the output spectrum of a system or device in response to a stimulus and is used to characterize the dynamics of the system. It is a measure of magnitude and phase of the output as a function of the input frequency. The first order of frequency response is known as the time constant. In magnetic field, there also exists another hysteresis. Fig. 5 is a magnetization curve that depicts the magnetization M of magnetic material with respect to the magnetic field H. The change in magnetization M when the magnetic field H is added to the degaussed state (H ¼ 0, M ¼ 0) and increased is called the initial magnetization curve. Magnetization increases along this curve and eventually saturates. Once saturated, if the magnetic field is reduced, it will not return to its original state, and a loop as shown by the arrow in the figure will be drawn. Hysteresis is a phenomenon in which loops occur due to different paths when the external field is changed from plus to minus and from minus to plus. If there is a hysteresis loop, it has two magnetization states, positive and negative, when the magnetic field is 0. Therefore, if these two values correspond to 1 and 0, non-volatile magnetic recording can be performed. In mechanical systems, there also is hysteresis. When the two gears are engaged, as shown in Fig. 6, the upper gear will follow when turning the lower gear to the left. However, when turning in the opposite direction, the rotation is not transmitted to the upper gear unless the lower gear rotates by the backlash angle. In this case, there are two stable states, one in which the lower gear is attached to the right wall of the upper gear and the other in which the lower gear is attached to the left wall, and hysteresis occurs because there is a threshold value of backlash in the response.

Temperature independence In some sensors, a long time affects the target range regardless of the measured value. This effect is known as drift. This is the lowfrequency change in a sensor with time. It is often associated with the electronic aging of components or reference standards in the sensor. Drift generally decreases with the age of a sensor as the component parts mature. A smoothly drifting sensor can be corrected for drift. For example, the gyroscope is a sensor that has a drift. The gyroscope drift is mainly due to the integration of two components: a slow-changing, near-DC variable called bias instability and a higher frequency noise variable called angular random walk (ARW). These parameters are measured in degrees of rotation per unit of time. The yaw axis is most sensitive to this drift.

Fig. 4 An alcohol thermometer. To change the volume, the temperature uses some quantity of heat. The decrease of the heat quantity induces the decrease of target temperature.

Physical Sensors: Mechanical Sensors

Fig. 5

67

A magnetization curve that depicts the magnetization M of magnetic material with respect to the magnetic field H.

What kind of mechanical sensors are needed for measuring contact state? Here, we consider the mechanical sensors for contact state. This is because contacting behavior is a very complex phenomenon for its bidirectionality, area of contact, and a combination of multiple physical information. Conventional pressure/force/thermal sensors measure information by using some physical laws. These laws are mainly linked to electronic signals. Some of them use a resistance shift, and the others use the electromotive force shift. This is because the signals are easily picked up by the electronic signals and integrated with other actuators. A few of them use optical fibers for the safety of the electric free system and the accuracy of measuring. The sensors which use electronic signals have amplification problem in itself. This is because the acquired original electronic signals are often small, and the S/N ratio may be a problem. The optical systems are free from these electronic amplification problems in itself. Though the sensors are useful for the tactile sensor in part, the sensors are not designed for the tactile sensor. So there are some defects for the tactile sensor.

Contact sensors of human Many sensors capture touch information. Humans also capture this rich information. Here, we will check what kind of sensory organ the human haptic sensation is. The haptic sense is a general term for sensory organs other than visual sense, auditory sense, taste sense, and olfactory sense, and includes cutaneous sense, proprioception, and equilibrium. The proprioception is the sensation of perceiving one’s joint angle and applied force by the Ruffini ending of the joint capsule, the Golgi tendon organ muscle of the joint ligament, the muscle spindle of the tendon, and the Golgi tendon organ. The sense of balance is acquired by semicircular canals and otoliths, which are in the inner ear. In addition, skin sensation of mechanical deformation is perceived by the following mechanoreceptors shown below. The cross-section of the hairless part of human skin has the structure shown in Fig. 7 (Kandel et al., 2000), and there are four types of mechanoreceptors (Merkel disk (SA I), Meissner’s corpuscle (FA I), Pacinian corpuscle (FA II), and Ruffini ending (SA II)). Using these receptors, humans use low frequencies to high frequencies of about 200 Hz for tactile recognition. And the high-frequency components mainly contain information on frictional sensations such as stick-slip.

Fig. 6 Backlash: The upper gear will follow when turning the lower gear to the left. However, when turning in the opposite direction, the rotation is not transmitted to the upper gear unless the lower gear rotates by the backlash angle. Adapted from Wikipedia.

68

Physical Sensors: Mechanical Sensors

Epidermis

Meissner' s corpuscle

Merkel disk

Dermis Pacinian corpuscle

Fig. 7

Structure of skin. Adapted from Kandel ER, Schwartz JH and Jessell TM (eds.) (2000). Principle of Neural Science. Appleton & Lange.

Bidirectionality Gibson mentioned that tactile sensation could be called Active Touch (Gibson, 1962). The tactile sensation is a relative sensory organ, and humans’ actions to touch objects themselves also indispensable elements to compose tactile information in the brain. In other words, unless the sensor itself has the softness of human skin, physical characteristics such as frictional characteristics, and reproduce tactile movements like humans, it cannot accurately acquire the same tactile information as humans obtain. As a discussion that summarizes the ideal requirements for human tactile sensors, there is a discussion of complete tactile sensors such as Harmon (Harmon, 1982). The following lists (Table 2) are a summary of the conditions required for a sensor to reproduce the human tactile sensation perfectly. Thus, it turns out that many conditions are required to aim for reproducing the human tactile sensation completely. To achieve human-like haptic sense as mechanical sensors, we should not consider the sensor based on some physical principles but the sensor design based on the required functions. There are at least three important points, spatial distribution, frequency distribution, and information transfer problem.

Spatial distribution The spatial distribution means that the tactile sensation has two-dimensional sensing distributions. The tactile itself is a boundary between humans and the environment. Based on simple topology, the boundary of a human whose body has three dimensions must be two-dimensional distributions. Many conventional sensors only measure the force/temperature of one point. This is because conventional sensors are not developed for the tactile sensor, and the applications often require only one point sensing information. However, in order to detect the changes of environment toward human, position information is also very much important. If there is some large pressure/thermal change information without a position, the human can detect hazardous information but cannot understand which way he/she should escape. In a smaller range, if there is rubbing movement of some object on the finger, the human can detect some time-varying information without position information. Though he/she can detect the changing information, he/she cannot detect the direction of the movement of the object. Without spatial information, the human cannot detect the changing direction of the signals. Of course, some distributed tactile sensor exists. For example, Pliance (Novel Corp, 2020), Tactilus seat type sensing (Sensor Products Inc., 2020), and Flexi force (Tekscan Inc, 2020). These are composed of small sensors unit and arranged in twodimensional arrays. They can measure the distribution of added forces. Each unit is independently connected. To analyze the contact state by many methods from the input data, the independence of the signal is useful. If we use these devices for tactile

Table 2

Requirements of human-like haptics sense.

Requirement

Reason

1 mm or less spatial frequency 1 kHz and above temporal frequency Measurement accuracy is 16 bits or more 3D vector distribution of force Elasticity Friction characteristics Low Hysteresis Robustness

Based on receptor distribution Based on receptor response characteristics Based on quantification of 0.1 mm  several mm Based on receptor distribution Relativity of contact Relativity of contact Required Conditions for All Sensors Requirements for all sensors

Physical Sensors: Mechanical Sensors

69

sensing applications, the independence may cause some problems. In the creating phase, the number of wirings may be a problem. In the measuring phase, because of the independence there is no network for signal processing. In order to acquire the position or movement information, we have to integrate and analyze the information after acquiring the input.

Range, dimension, and material The force distribution means that the tactile sensor should measure the required force range and force vectors. Some sensors have their specific sensing ranges owing to their physical law. If the physical law define the range that is different from humans’ one, we have to prepare many sensing devices for the sufficient human sensing range. So the sensing method itself should not define the range. The ideal sensor should have enough range for humans. However, there is still no such sensor in the world. The second best sensor is a range changeable sensor by designing the sensing element without changing the sensing method. The force vector information is also important. A force (F)/torque (T) toward one point has three-dimensional components, Fx, Fy, Fz, Tx, Ty, and Tz. Some conventional sensors can measure such information on only one point, but the spatial and force distribution coincides. So the ideal sensor should measure the distribution of force vectors. Additionally, the sensor itself should have near or the same characteristics of the material as the humans’. For example, Young’s modulus, Poisson’s ratio, and friction coefficient are essential aspects of the sensed information. This is because the stress and the strain are indivisible. So the sensing information between by strained sensing surface and by unstrained sensing surface is different. Furthermore, the friction coefficient also should be the same between the sensor and human. The essential thing in tactile sensing is that the reproduction of the same contact state between the contact by the sensor and the contact by human skin.

Some devices for force distribution Here Kamiyama et al. proposed a tactile sensor called GelForce (Kamiyama et al., 2005). The sensor is made of silicone rubber and an imaging device. Inside the silicone rubber, there are two-layered marker patterns. By capturing the displacement of the markers with the imaging device, it can reconstruct the measured information. In the previous section, the imaging device has well-switching technology, so the use of the imaging device is proper for two-dimensional sensing. The sensor can measure the force and spatial distribution simultaneously. It measures two-dimensional distributions of three-dimensional forces (Fx, Fy, Fz) and threedimensional torques (Tx, Ty, Tz). Furthermore, the simplicity of the component of the sensor with silicone rubber, the responsibility can be changed easily. With this feature, it realizes almost the same characteristics of the material as humans’. We also research a similar sensor, named reflection-type tactile sensor, with the use of silicone rubber and an imaging device (Saga et al., 2007) (Fig. 8). The sensor can measure the displacement of the sensor surface by using the total reflection of the contact

Fig. 8

Reflection type tactile sensor.

70

Physical Sensors: Mechanical Sensors

surface. According to the use of the reflection image, the sensor realizes high-resolution sensitivity and can detect 0.01 mm displacement of the sensor surface. The same with the GelForce, the sensor can design the characteristics of the material. By controlling the image pattern dynamically, the sensor reconstructs the surface deformation more precisely (Saga et al., 2014). To achieve more precise reconstruction and to handle large deformation, we considered the implementation of active patterns for the input image patterns. This method has a characteristic of zero method. We suggested that the optimum image pattern was that whose reflection image had an equal distribution of reflection points, and we proposed a method of successive active pattern generation accordingly. The method uses the estimated reconstructed shape of the previous frame to generate active patterns whose reflection points are distributed evenly. By employing the reflection image generated by the active patterns, the reconstruction continues successively.

Frequency distribution The frequency distribution means that, from the Fig. 2, humans’ receptors have their own responsiveness toward the frequency domain. For example, Merkel disks (SA I) have their responsibility about 5–15 Hz, Meissner’s corpuscles (FA I) have about 20– 50 Hz, and Pacinian corpuscles (FA II) have about 60–400 Hz. From this fact, the tactile sensor should measure at least the range of 0–400 Hz. Many kinds of conventional tactile sensors have such a measurement range (Fig. 9; Gesheider et al., 2001). So the frequency distribution seems sufficient for the tactile sensor. This is because the sensor often measures only one point. Sampling with a few points will make the responsiveness of the sensor faster.

Information transfer problem If we treat the frequency distribution and the spatial distribution at the same time, the frequency distribution becomes difficult problem to solve. Though the sensors’ ability toward the frequency distribution is high, the multiple sensors arranged in twodimensional arrays require data collecting methods, such as scanning or matrix switching. Methods of acquiring twodimensional discrete data are critical for frequency distribution. Simple scanning method requires n  n ordered wirings and n  n switching device. Matrix switching method requires 2n ordered wirings and n  n switching device. Each method requires n  n ordered scanning speed. This is because each method aims to get all of the acquired information.

Imaging devices for information transfer Here, the imaging device also has such switching technology, CMOS (Complementary Metal Oxide Semiconductor) imaging sensor and CCD (Charge-coupled device) imaging sensor. CMOS imaging sensor is known as an active-pixel sensor (APS), also commonly written active pixel sensor. It is an image sensor consisting of an integrated circuit containing an array of pixel sensors, each pixel containing a photodetector and an active amplifier. Many types of active pixel sensors include the CMOS imaging sensor used most commonly in web cameras. A CMOS process produces this imaging sensor, so it is also known as a CMOS imaging sensor. Because of its simplicity, the CMOS imaging device can realize block scanning. By separating the imaging area into some blocks, the device can scan each block simultaneously. In recent years Sony Inc. creates fast scanning chips by specialized design (Barth et al., 2007). CCD itself is an analog shift register, enabling electric charges to be transported through successive capacitors controlled by a clock signal. A CCD can be used as a form of memory or for delaying analog, sampled signals. By using this device, a CCD imaging sensor is created for serializing parallel analog signals.

2.9 cm² stimuli 0.008 cm² stimuli P ch. (FA II) NP I ch. (FA I)

NP II ch. (SA II) NP III ch. (SA I)

0.1

1

10

100

1000

Fig. 9 Responsiveness of the receptors. Adapted from Gesheider GA, Bolanowski SJ and Hardick KR (2001) The frequency selectivity of information-processing channels in the tactile sensory system. Somatosensory and Motor Research 18(3):191–201.

Physical Sensors: Mechanical Sensors

71

Other novel devices for information transfer Another communicating device, Two-Dimensional Signal Transmission (2DST), is developed by Shinoda et al. (2007). This device realizes the communication between each element without wiring them independently. The device is made from some layered conductive sheets, and by using microwave confined around the surface, it realizes the low power and high-security communication. This technology is designed for the use of tactile information and now developing. By using the special transmitting protocol, this device may realize high-speed transmission of information and compression technology. Some methods are compressing sensing information without losing important aspects of tactile information. The soft tribo-sensor using PolyVinylidene DiFluoride (PVDF) Film is created by Jiang et al. (1999) Though this sensor has only a one-dimensional measurement point, it can measure high-frequency pressure change. By scanning the sensor itself on skin surfaces, it can measure the difference of them. Another thin and flexible tactile sensor is made of ordinal pressure-conductive rubber, though the wirings of the sensor are very few (Shimojo and Ishikawa, 1990). Furthermore, the sensor itself is a flexible sheet. These sensors have been used as skins of some robots. The sensor’s measurement information is limited only to the center of mass and the mass itself. By limiting the information, the sensor requires only four wirings for each area. For example, manipulation of some object with a robot arm requires only this information.

Is every distribution required? Though each distribution still cannot be combined now, the combination of these ranges will open the new sensing features (pain, itchy, tickle, and feel good) for the tactile sensor. Additionally, the important aspect of the tactile sensation is as follows. However, we cannot compose a perfect sensor with these distributions. We should well consider the application of the sensor and design it. The application decides the sensor’s required information, so the limitation of information matters little.

Deep learning for haptic information In this section, we discuss machine learning technologies for sensing.

Classification of haptic information using deep learning Nowadays, these technologies have been positively developing. There are many well-known open image databases in the computer vision field where the technologies have been sophisticated, such as NORB, Caltech-101/256, and CIFAR-10/100. Deep Learning, a method of machine learning, has come into the limelight due to image recognition contests using such databases. The development of these machine learning technologies is also benefiting tactile sensing (Strese et al., 2017; Abdulali and Jeon, 2016). Strese et al. (2017) implemented 108 texture classifications under some conditions using six types of data such as acceleration, pressure, temperature, image, sound, and magnetic field acquired by a pen-type device. As a result of classification by Euclidean distance, Mahalanobis distance, and mixed Gaussian model, the average conformity rate was around 85% as a whole. Abdulali and Jeon (2016) used an acceleration sensor and a force sensor attached to a haptic interface, and by operating them in two dimensions, the position, velocity, acceleration, and force are acquired at the same time. Then they implemented a system that classifies each signal using the Radial Basis Function Network. Burka et al. (2016) proposed a device that collects visual and tactile information on the surface of an object for the purpose of application to autonomous robots. Jamali and Sammut (2010) proposed a method for collecting data on the surface of an object using a device with a piezoelectric film built into an artificial finger and collected information on the surface of seven types of objects. Note that these systems collect each information as tactile information as texturespecific information and incorporate human tactile movements for tactile perception into machine learning. In this way, a method of recording and classifying tactile information is being realized by using some sensor inputs in combination with machine learning technology. In the next section, we will introduce our efforts on a method to realize the classification with fewer sensors while incorporating such human tactile movements.

Acquisition of tactile information and their classification using machine learning As mentioned in the previous section, in recent years, a method of recording and classifying tactile information has been realized by using some sensor inputs in combination with machine learning technology. We also consider that it would be possible to realize the same classification with fewer sensors while incorporating such human tactile movements, and proposed a tactile information classification system using only acceleration sensors (Agatsuma et al., 2020). Normally, the behavior that humans trace to identify textures and the friction between the texture and the fingers causes vibrations. The accelerometer attached near the texture captures information, including both this active motion and the vibration generated from the texture. Therefore, we focused on this acceleration and proposed a method of constructing a wireless sensor system that can be attached near the object, and collecting and classifying the acceleration information. In this approach, we collect, classify, and generate only acceleration as tactile information. By using only acceleration data, we collect the information easier than conventional research. Furthermore, employing a machine learning-based classification and generation method, we propose a consistent handling approach of the information for tactile displays. By using the ZigBee-

72

Physical Sensors: Mechanical Sensors

based microcomputers and implemented a Convolutional Neural Network (CNN)-based classification method of haptic information (Fig. 10), we succeeded in classifying 30 types of data with an accuracy of about 88.9% (Table 3). In robotics field, Sundaram et al. (2019) revealed the human grasp’s signatures using a scalable tactile glove. Using a low-cost scalable tactile glove sensor array, they record a large-scale tactile dataset with 135,000 frames, each covering the full hand while interacting with 26 different objects. This set of interactions with different objects reveals the critical correspondences between different regions of a human hand while it is manipulating objects.

Classification as a function of sensors These classification methods separate the measured value into several discrete classes. Normally, sensors should respond to the measured data. However, the previous classification method also meets the sensor function. In most CNN classification models, the output layer is the softmax layer. This is because the classifier should output only one hot vector for classification. However, by employing linear regression in machine learning, we can estimate sequential values. It is very similar to the use of deep learning for the classification problem. You can use different layers at the end of the network. e.g., in CNN, instead of a softmax layer and cross-entropy loss, you can use a regression layer and mean-square error loss, etc. With the layer, you can construct a sensing function by using regression output from multiple input information.

Generation of sensing information by deep learning As we have introduced so far, some researchers have used machine learning based on collected data to classify the textures (Abdulali and Jeon, 2016; Strese et al., 2017; Agatsuma et al., 2018). However, there are enormous amounts of conditions of the combination between textures and stroking motions, and these studies cannot cover all of these combinations. For example, Strese et al. (2017) collected data on various conditions with a pen-type device, though an acute angle against the object is fixed. Thus the data that can be obtained may change if the angle is changed. In this way, it is unrealistic to comprehensively collect data from an actual object because of the significant number of conditions to consider. The existence of data of conditions that cannot be collected also means that tactile information of objects under those

Fig. 10

Left: Composition of our CNN model, Right: 15 textures. These are planer object with length 70–100 mm and width 100–130 mm.

Table 3

Confusion matrix of 30 kinds of data classification under 400 mm/s movement. B

C

D

E

F

G

H

I

J

K

L

M

N

O

P

Q

R

S

T

U

V

W

X

Y

Z

AA

AB

AC

AD

0.94 0 0 0 0.02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0.92 0 0 0 0.04 0 0 0 0 0 0.04 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0.92 0 0.02 0 0 0 0 0 0.01 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0.97 0 0.03 0 0 0 0 0 0.03 0 0 0 0.01 0 0 0 0 0.01 0 0 0 0 0 0 0 0 0

0.02 0 0.05 0 0.95 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0.03 0 0 0 0.9 0 0 0 0 0 0 0 0.01 0 0.01 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0.03 0 0 0 0 0 0.96 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0.01 0 0.92 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0 0 0 0 0.01

0 0 0.02 0 0 0 0.01 0 0.98 0 0.03 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0.01 0 0 0 0 0.01 0 0 0 0.94 0 0 0 0.01 0 0 0 0.02 0 0 0 0 0 0.02 0 0 0 0 0 0

0 0 0 0 0.01 0 0.01 0 0.02 0 0.95 0 0 0 0 0 0 0 0.01 0 0 0 0 0 0 0 0 0 0 0

0 0.01 0 0 0 0.02 0 0 0 0.03 0 0.91 0 0.01 0 0 0 0.01 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0.01 0 0 0 0 0 0.98 0 0.01 0 0.01 0 0 0.01 0 0 0 0 0.01 0.01 0 0 0 0

0 0 0 0 0 0 0.01 0 0 0 0.01 0.01 0.01 0.95 0 0 0 0 0 0 0 0 0 0.04 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.86 0 0 0 0 0.03 0 0 0 0 0 0 0 0 0.08 0

0 0.01 0 0 0 0 0 0 0 0.01 0 0 0 0 0.01 0.92 0 0.01 0 0 0.01 0 0 0 0 0.01 0.04 0 0.01 0

0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0.02 0 0.9 0 0.02 0.01 0 0 0 0 0 0.01 0 0 0.02 0

0 0 0 0 0 0 0 0 0 0.03 0 0 0 0 0 0.03 0 0.94 0 0 0.01 0 0 0 0 0 0.01 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0.87 0 0 0.02 0 0 0.01 0 0 0 0 0

0 0 0.01 0 0 0 0 0 0 0 0 0 0 0 0.02 0 0.01 0 0.01 0.85 0 0 0 0 0 0.01 0 0 0.03 0

0 0.01 0 0.02 0 0 0 0 0 0 0 0.01 0 0 0 0 0 0 0 0 0.97 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0 0 0.02 0.01 0 0.95 0 0 0 0 0 0 0.01 0

0 0 0 0 0 0 0 0.05 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.99 0 0 0 0 0.01 0 0.02

0 0.01 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0 0 0 0 0 0 0.89 0 0 0 0 0 0.03

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0.03 0 0 0 0 0.01 0.97 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.04 0.03 0 0.01 0 0.02 0 0.94 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0.01 0 0.02 0 0.02 0 0 0 0 0 0.01 0 0 0.94 0 0 0

0 0 0 0 0 0 0 0.02 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.99 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.08 0 0.07 0 0.01 0.06 0 0.01 0 0 0.01 0.01 0 0 0.86 0

0 0 0 0.01 0 0 0 0.01 0 0 0 0 0 0.01 0 0 0 0 0 0 0 0 0.01 0.01 0 0 0 0 0 0.94

Physical Sensors: Mechanical Sensors

A: carpet1-pen B: carpet1 C: carpet2-pen D: carpet2 E: carpet3-pen F: carpet3 G: sponge-b-pen H: sponge-b I: sponge-g-pen J: sponge-g K: sponge-y-pen L: sponge-y M: stonetile1-pen N: stonetile1 O: stontile2-pen P: stonetile2 Q: stonetile3-pen R: stonetile3 S: whitetile1-pen T: whitetile1 U: whitetile2-pen V: whitetile2 W: whitetile3-pen X: whitetile3 Y: woodtile1-pen Z: woodtile1 AA: woodtile2-pen AB: woodtile2 AC: woodtile3-pen AD: woodtile3

A

73

74

Physical Sensors: Mechanical Sensors

conditions cannot be classified. Therefore, we consider this problem and propose a new method to replace the method of collecting data directly from an object. Instead of an exhaustive direct data collection method, machine learning is used to generate alternative data from the minimum amount of collected data (Agatsuma et al., 2020). As a result, the data collection cost can be minimized. Besides, by adjusting the machine learning model, it is possible to generate data that will replace the data that has not been collected. As a first step in realizing the proposed method, we implemented a data generation method based on machine learning that focuses on the stroking movement’s acceleration data. Here, we consider the role of GAN from the sensing viewpoint. The regression layer in the CNN model can evaluate the information which is in the input data. However, if there is no data for evaluation in the input data, the model cannot evaluate the data. Here, we evaluate the data without any input data for evaluation by using the GAN model. Nowadays, the intuitive index, such as the texture, the moisture feelings, was indexed by sensory evaluation. However, the indexes are different between participants. Here, with the GAN model, we evaluate the unindexed data such as textures, moisture feelings, or so. By using the evaluation, we can parameterize and sense the intuitive index with a more systematic method.

Conclusion In this article, we have introduced and reviewed principles, functionality, and future of mechanical sensors. First, we have considered what the mechanical sensors are. Mechanical sensors measure the displacement/movement/direction/acceleration of measured target. The measurement methods vary, though we have picked up the sensors which measure the mechanical changes. First, we have reviewed the mechanical sensors from a conventional viewpoint, mathematical viewpoint, and haptic information viewpoint. Conventional contact sensors have been created from some principles of physics. They record the phenomena of the contact status using some physical principles. Further, sensor is like function in mathematics, which is a binary relation over two sets that associates every element of the first set (the domain of definition, domain), to exactly one element of the second set (range). The domain values are measured as raw data, and the range values are target values indicated by the measured raw data. The critical points are the independence of measurement, target value history, and temperature. Many sensors capture touch information. Humans also capture this rich information. Then, we have checked what kind of sensory organ the human haptic sensation is. The essential points are its bidirectionality, spatial distribution, frequency distribution, and information transfer. After these reviews, we have discussed the potential of machine learning as the future of mechanical sensors. The regression layer in the CNN model can evaluate the information which is in the input data. However, if there is no data for evaluation in the input data, the model cannot evaluate the data. Here, we evaluate the data without any input data for evaluation by using the GAN model. By using the GAN model, we can parameterize and sense the intuitive index with a more systematic method. For further multifunctional mechanical sensing, the machine learning method will be essential technology.

References Abdulali, A., Jeon, S., 2016. Data-driven modeling of anisotropic haptic textures: Data segmentation and interpolation. In: International Conference on Human Haptic Sensing and Touch Enabled Computer ApplicationsSpringer, pp. 228–239. Agatsuma, S., Nakagawa, S., Ono, T., Saga, S., Vasilache, S., Takahashi, S., 2018. Classification method of rubbing haptic information using convolutional neural network. In: Proceedings of International Conference, HCI International 2018, pp. 159–167. Agatsuma, S., Kurogi, J., Saga, S., Vasilache, S., Takahashi, S., 2020. Simple Generative Adversarial Network to Generate Three-Axis Time-Series Data for Vibrotactile Displays. In: Proceedings of International Conference on Advances in Computer-Human Interactions, ACHI 2020. Ando, S., Shinoda, H., Yonenaga, A., Terao, J., 2001. Ultrasonic six-axis de- formation sensing. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control 48 (4), 1031–1045. Barth, J., Reohr, W., Parries, P., Fredeman, G., Golz, J., Schuster, S., Matick, R., Hunter, H., Tanner, C., Harig, J., Kim, H., Khan, B., Griesemer, J., Havreluk, R., Yanagisawa, K., Kirihata, T., Iyer, S., 2007. A 500MHz random cycle 1.5ns-latency, soi embedded dram macro featuring a 3T micro sense amplifier. In: Digest of Technical Papers of IEEE International Solid-State Circuits Conference, pp. 486–617. Benes, E., Groschl, M., Seifert, F., Pohl, A., 1998. Comparison between baw and saw sensor principles. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control 45 (5), 1314–1330. BL Autotec Ltd, 2020. BL Nano Sensor. https://www.bl-autotec.co.jp/english/. Accessed on 25 Nov 2020. Burka, A., Hu, S., Helgeson, S., Krishnan, S., Gao, Y., Hendricks, L.A., Darrell, T., Kuchenbecker, K.J., 2016. Design and implementation of a visuo-haptic data acquisition system for robotic learning of surface properties. In: Proc. IEEE Haptics Symposium. Gesheider, G.A., Bolanowski, S.J., Hardick, K.R., 2001. The frequency selectivity of information-processing channels in the tactile sensory system. Somatosensory and Motor Research 18 (3), 191–201. Gibson, J.J., 1962. Observations on active touch. Psychological Review 69, 477–491. Harmon, L.D., 1982. Automated tactile sensing. International Journal of Robotics Research 1 (2), 3–32. Jamali, N., Sammut, C., 2010. Material classification by tactile sensing using surface textures. In: 2010 IEEE International Conference on Robotics and AutomationIEEE, pp. 2336–2341. Jamone, L., Natale, L., Metta, G., Sandini, G., 2015. Highly sensitive soft tactile sensors for an anthropomorphic robotic hand. IEEE Sensors Journal 15 (8), 4226–4233. Jiang, Z., Funai, K., Tanaka, M., Chonan, S., 1999. Development of soft tribo-sensor using PVDF film for skin surface contour measurement. Journal of Intelligent Material Systems and Structures 10, 481–488. Johnson, M.K., Adelson, E.H., 2009. Retrographic sensing for the measurement of surface texture and shape. In: Computer Vision and Pattern Recognition (CVPR), pp. 1070–1077. Kamiyama, K., Vlack, K., Kajimoto, H., Kawakami, N., Tachi, S., 2005. Vision-based sensor for real-time measuring of surface traction fields. IEEE Computer Graphics & Applications Magazine 25 (1), 68–75. Kandel, E.R., Schwartz, J.H., Jessell, T.M. (Eds.), 2000. Principle of Neural Science. Appleton & Lange.

Physical Sensors: Mechanical Sensors

75

Lee, H.-K., Chang, S.-I., Yoon, E., 2009. Dual-mode capacitive proximity sensor for robot application: Implementation of tactile and proximity sensing capability on a single polymer platform using shared electrodes. IEEE Sensors Journal 9 (12), 1748–1755. Maekawa, H., Tanie, K., Komoriya, K., 1993. A finger-shaped tactile sensor using an optical waveguide. In: Proceedings of IEEE Systems Man and Cybernetics Conference-SMC, vol. 5. IEEE, pp. 403–408. Maldonado, A., Alvarez, H., Beetz, M., 2012. Improving robot manipulation through fingertip perception. In: 2012 IEEE/RSJ International Conference on Intelligent Robots and SystemsIEEE, pp. 2947–2954. Medical Tactile Inc, 2020. Digitact. https://pressureprofile.com/. Accessed on 25 Nov 2020. Novel Corp. Pliance (2020) (Accessed on 25 Nov 2020). https://www.novel.de/products/pliance/ OptoForce Inc, 2020. Hex Force/Torque Sensor. https://optoforce.com/. Accessed on 25 Nov 2020. Saga, S., 2008. How tactile sensors should be? chapter 1 In-Tech, pp. 1–14. Saga, S., Kajimoto, H., Tachi, S., Feb. 2007. High-resolution tactile sensor using the deformation of a reflection image. Sensor Review 27, 35–42. Saga, S., Taira, R., Deguchi, K., 2014. Precise shape reconstruction by active pattern in total-internal-reflection-based tactile sensor. IEEE Transactions on Haptics 7 (1), 67–77. Schmitz, A., Maggiali, M., Natale, L., Bonino, B., Metta, G., 2010. A tactile sensor for the fingertips of the humanoid robot icub. In: 2010 IEEE/RSJ International Conference on Intelligent Robots and SystemsIEEE, pp. 2212–2217. Sensor Products Inc, 2020. Tactilus. http://tactilus.net/. Accessed on 25 Nov 2020. Shimojo, M., Ishikawa, M., 1990. Thin and flexible position sensor. Journal of Robotics and Mechatronics 2 (1), 38–41. Shimojo, M., Namiki, A., Ishikawa, M., Makino, R., Mabuchi, K., 2004. A tactile sensor sheet using pressure conductive rubber with electrical-wires stitched method. IEEE Sensors Journal 4 (5), 589–596. Shinoda, H., Makino, Y., Yamahira, N., Itai, H., 2007. Surface sensor network using inductive signal transmission layer. In: Proceedings of 4th International Conference on Networked Sensing Systems, Braunschweig, Germany. Strese, M., Boeck, Y., Steinbach, E., 2017. Content-based surface material retrieval. In: World Haptics Conference (WHC), 2017 IEEEIEEE, pp. 352–357. Sundaram, S., Kellnhofer, P., Li, Y., Zhu, J.-Y., Torralba, A., Matusik, W., 2019. Learning the signatures of the human grasp using a scalable tactile glove. Nature 569 (7758), 698–702. Tekscan Inc, 2020. Flexi Force. https://www.tekscan.com/products-solutions/embedded-force-sensors. Accessed on 25 Nov 2020. Terada, K., Suzuki, Y., Hasegawa, H., Sone, S., Ming, A., Ishikawa, M., Shimojo, M., 2011. Development of omni-directional and fast-responsive net- structure proximity sensor. In: 2011 IEEE/RSJ International Conference on Intelligent Robots and SystemsIEEE, pp. 1954–1961. Tremblay, M.R., Cutkosky, M.R., 1993. Estimating friction using incipient slip sensing during a manipulation task. In: [1993] Proceedings IEEE International Conference on Robotics and Automation. IEEE, pp. 429–434. Yan, J., 2014. Machinery Prognostics and Prognosis Oriented Maintenance Management. John Wiley & Sons.

Physical Sensors: Acoustic Sensors Osamu Saito, Fengming Yu, and Yoji Okabe, Department of Mechanical and Biofunctional Systems, Institute of Industrial Science, The University of Tokyo, Tokyo, Japan © 2023 Elsevier Ltd. All rights reserved.

Introduction Piezoelectric sensor Piezoelectric effect Electroacoustic conversion Ferroelectrics Representative piezoelectric materials Sensors based on piezoelectricity General components of piezoelectric sensor Vibration modes Resonance Macro fiber composite Piezoelectric sensing at high temperatures Loss of piezoelectricity at high temperatures Piezoelectric materials adaptable for high temperatures Buffer rod method Fiber optic sensor Fiber-optic interferometry-based acoustic sensing Fiber Bragg grating sensor Fiber Bragg grating Bragg grating spectra Acoustic measurement with FBG sensor Phase-shifted FBG for highly sensitive acoustic measurement Ultrasonic measurement at high temperature by PSFBG sensor Remote adhesion configuration for acoustic emission measurement Regeneration of Bragg grating Conclusions References Relevant Websites

76 77 77 77 78 79 80 80 80 81 83 83 84 85 85 86 86 87 87 87 88 89 90 91 92 94 94 96

Abstract Acoustic wave sensors have been widely used in various industries. In this chapter, we provide an introductory guide to sensors based on piezoelectricity and optical fiber sensors such as a fiber Bragg grating sensor. This chapter also presents the application of these sensors to a high-temperature environment because the demand for this application has recently increased.

Introduction Acoustic waves (or ultrasonic waves) are used for many purposes in various industries. For example, nondestructive evaluation based on ultrasonic waves is conducted in the aerospace, automobile, and civil engineering industries to detect structural defects (Mix, 2005). Structural health monitoring (SHM) in the abovementioned industries also requires sensing acoustic waves (Giurgiutiu, 2014). In addition, the observation of acoustic waves is the basis of sound navigation and ranging (SONAR) devices, which determine the position of fish in ocean engineering (Urick, 2013) and diagnostic equipment for human health in medical engineering (Jensen, 2007). Thus, acoustic wave sensors have undergone development over a long time. Recently, the demand for monitoring systems in harsh environments, such as high-temperature environments, has increased. For example, monitoring the internal combustion engines of airplanes and vessels enables the improvement in combustion efficiency and reduces exhaust gas (Zu et al., 2016). Moreover, continuous monitoring of thermal power plant and geothermal plant equipment can ensure plant safety and a stable electricity supply. The data acquired by acoustic sensors can be used for anomaly detection through big data analysis and machine learning, which are developing rapidly (Chandola et al., 2009).

76

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00067-4

Physical Sensors: Acoustic Sensors

77

Piezoelectric and optical fiber sensors are known as acoustic wave sensors. The former are based on the piezoelectric effect, which generates electricity by application of stress, while the latter are based on optical properties: interference and reflectivity change; extensive research has focused on these sensors. This chapter introduces piezoelectric and fiber optic sensors and serves as a beginner’s guide to acoustic sensing. Hence, we review elemental aspects from many textbooks and reviews. This chapter also introduces acoustic measurements at high temperatures because sensors applicable to high-temperature structures have been in demand. In sections “Piezoelectric sensor” and “Fiber optic sensor,” piezoelectric sensors and fiber optic sensors are discussed, respectively, and the final section concludes the chapter.

Piezoelectric sensor This section describes acoustic sensors based on piezoelectricity. First, we introduce the piezoelectric effect and then consider its application to sensing. Finally, we discuss the applicability of piezoelectric sensors to high-temperature environments.

Piezoelectric effect Electroacoustic conversion Ordinary sensors for acoustic waves are based on the piezoelectric effect through which electric charges are induced in response to applied stress. Well-known materials that exhibit the piezoelectric effect include quartz and Rochelle salt (Valasek, 1921). When stress is imposed on piezoelectric materials by incident acoustic waves, they produce electric charges. Therefore, by measuring the resultant electric signals, we can detect the incident acoustic waves. In contrast to the piezoelectric effect, if we apply an electric field to piezoelectric materials, they exhibit expansion (or contraction). This is known as the converse piezoelectric effect. By applying a voltage to piezoelectric materials, we can generate acoustic waves in contact materials; thus, piezoelectric materials can be used as actuators as well as sensors. Physical variables related to the piezoelectric effect are shown in Fig. 1. These include strain SI (I ¼ 1, 2, ., 6) and stress TJ (J ¼ 1, 2, ., 6) as mechanical variables and electrical displacement Di and electric field Ei (i ¼ 1, 2, 3) as electrical variables; we use Voigt’s notation for mechanical variables. In our notation, uppercase indices run from 1 to 6, whereas lowercase indices run from 1 to 3. These variables uphold the following piezoelectric fundamental equations (Ballantine et al., 1996): SI ¼ sEIJ TJ þ d0kI Ek

(1)

Di ¼ diJ TJ þ 3 Tij Ej

(2)

and

E

where sij are elastic compliance-related mechanical variables. The superscript E indicates that the compliance is measured under the condition of a constant electric field E. Similarly, 3 ijT are dielectric constants under constant stress T. Mechanical and electric variables are coupled through the direct piezoelectric constant diJ of Eq. (2) and converse piezoelectric constant dkI0 of Eq. (1),  this  can be inferredas follows.  respectively. dkI0 is identical to dkI, and These variables can be expressed as second order differentials vG and d ¼  v vG . Since the order of the differentiation can be changed, the variables of the Gibbs energy G: d0kI ¼  vEv k vT kI vTI vEk I are the same. From the coefficients of Eqs. (1) and (2), the electromechanical coupling coefficients k2 can be defined, for example:

Fig. 1 Correlation between mechanical variables and electrical variables. Modified from Ballantine Jr DS, White RM, Martin SJ, et al. (1996) Acoustic Wave Sensors: Theory, Design, and Physico-chemical Applications, 1st edn., p. 23. Elsevier.

78

Physical Sensors: Acoustic Sensors

ðk31 Þ2 ¼

ðd31 Þ2 sE11 3 T33

(3)

for which the physical meaning is the conversion ratio of the input electrical energy to the mechanical energy, and vice versa. To understand this, suppose a case where the xx-component of strain S1 is induced by the application of the z-component of electric field E3 under the stress-free condition (T ¼ 0). The input electric energy can be calculated as 1 T 2 3 E 2 33 3

(4)

1 ðS1 Þ2 1 ðd31 E3 Þ2 ¼ : 2 sE11 2 sE11

(5)

Uelec ¼ and the stored mechanical energy as Umec ¼ Hence, we can show the following:

Umec ðd31 Þ2 ¼ E T ¼ ðk31 Þ2 : Uelec s11 3 33

(6)

Ferroelectrics We consider the origin of piezoelectricity. As a representative example, we investigate lead titanate PbTiO3, the crystal structure of which is known as a perovskite-type structure. The ideal perovskite-type structure is shown in Fig. 2A (Feynman et al., 1971; Uchino, 2015). Atoms of metals A and B are located at the eight vertices and the center of the unit cube, respectively. Metal B is surrounded by six O atoms, which are located at the faces of the unit cube and form an octahedron. The atoms of metals A and B are positively charged, whereas the O atoms are negatively charged. Fig. 2B shows the crystal structure of PbTiO3 in which metal A is replaced by Pb and metal B by Ti. Note that the locations of Pb and Ti are shifted against those of the O atoms. An electric dipole moment is generated owing to the positively charges Pb and Ti and the negatively charged O atoms. The abovementioned relative displacement of positive and negative charges is a mechanism that generates polarization. If we impose compressive stress along the polarization direction, the shift between positively charged and negatively charged atoms decreases; thus, the polarization is reduced. This is the origin of the piezoelectric effect. By contrast, if we apply an electric field along the dipole moment, the positively charged atoms move further toward the positive direction of the dipole moment, and negatively charged atoms move in the opposite direction. Hence, the crystal extends along the electric field. This is the converse piezoelectric effect. Further, we consider the piezoelectric material at the macroscopic level. As shown on the left in Fig. 3, piezoelectric materials are generally divided into many domains (Mitsui and Furuichi, 1953), each of which has its polarization direction. When the angle of the polarization changes by q degrees at the boundary between domains, the boundary is called a q domain wall (q can be 71 , 90 , 109 , and 180 , depending on the materials). Owing to the domain structures, the polarization does not appear in macro-scale materials due to cancelation by averaging, even if the polarization spontaneously occurs at microscopic levels. Strong electric fields are applied to piezoelectric materials to prepare them for use as sensors. In the vicinity of the domain walls, the polarization direction and electric field are changed. Thus, the domain walls move, and the domains polarized along the electric field expand, as shown on the right in Fig. 3. This polarization alignment by strong electric fields is called poling, through which piezoelectric materials can be practically utilized. Generally, materials that have a spontaneous polarization that can change direction in an external electric field are ferroelectrics, of which PbTiO3 and BaTiO3 (Ogawa, 1947; Vul, 1946; Wainer, 1946) are representative materials. Every ferroelectric material exhibits piezoelectricity (Jaffe et al., 2012).

Fig. 2 Appearance of piezoelectricity from perovskite-type structure. (A) Perovskite-type structure. (B) Electric dipole moment of PbTiO3. Modified from Tanaka T, Okazaki K, and Ichinose N (1973) Piezoelectric Ceramic Materials, 1st edn., p. 40. Gakukensya (in Japanese).

Physical Sensors: Acoustic Sensors

79

Fig. 3 Domain structure of BaTiO3. The arrows indicate the direction of polarization. Modified from Watanabe T (2020) Amazing piezoceramic PZT. Journal of the Japan Society for Precision Engineering 86: 213–216 (in Japanese).

Representative piezoelectric materials Many piezoelectric materials have been discovered and developed. Fig. 4 presents some of them. Types of piezoelectric materials include ceramics, single crystals, thin films, and polymers; ceramics are polycrystalline in contrast to single crystals. The production of single crystals is difficult; however, they are an attractive material because piezoelectricity and permittivity are dependent on crystal orientation, and we can select the optimal orientation for practical use. Piezoelectric thin films have also been applied to high-frequency sensing. Piezoelectric polymers are employed in various industries owing to their flexibility. The following discussion presents representative piezoelectric materials. Lead zirconate titanate (PZT) is a solid solution of lead zirconate (PbZrO3) and lead titanate (PbTiO3). Initially, the word “PZT” was the brand name of a product by Clevite Corporation. However, the abbreviation is now used for lead zirconate titanate (Pb(Zr, Ti)O3). Its properties are changed based on the mixing ratio of the two materials (Sawaguchi, 1953). When zirconate is higher than 52%, PZT is in the rhombohedral phase; however, a lower zirconate percentage indicates PZT in the tetragonal phase. The piezoelectric constant is significantly enhanced at the boundary between the two phases (Jaffe et al., 2012), which is called the morphotropic phase boundary. PZT is one of the most widespread piezoelectric materials owing to its high piezoelectric constants. Polyvinylidene fluoride (PVDF) is a piezoelectric polymer (Kawai, 1969) that is produced by the polymerization of vinylidene difluoride ([–CH2–CF2–]n), and it is known as a highly non-reactive thermoplastic. Although its piezoelectric constant is smaller than that of PZT, and the electromechanical coupling is less efficient (Eaton et al., 2009), PVDF sensors are used in various industries

Fig. 4 Representatives of piezoelectric materials. Modified from Honda Electronics Co., Ltd. World of Inaudible Sound, p. 215 (in Japanese). https:// f.hubspotusercontent20.net/hubfs/9272528/images/product/ufile/library/2359_file.pdf (Accessed 15 June 2021).

80

Physical Sensors: Acoustic Sensors (A)

(B)

Electric wires Backing material

Piezoelectric material

Electrodes Protective layer

Acoustic wave

Fig. 5

Object

General piezoelectric sensor. (A) Photograph of piezoelectric sensors (AE304S, Fujiceramics, Japan). (B) Structure of AE sensor.

owing to their advantages, such as flexibility (Xin et al., 2016). As an example, applications to SHM have been investigated (De Rosa and Sarasini, 2010; Lin and Giurgiutiu, 2006). Additionally, PVDF is also employed in medical engineering because the ultrasonic transmittance between PVDF and human bodies is high (owing to the matching of acoustic impedances). Relaxer ferroelectrics have been investigated to obtain materials with stronger piezoelectric properties than those of PZT; these include lead magnesium niobate (Pb(Mg1/3Nb2/3)O3: PMN) and lead-zinc niobate (Pb(Zn1/3Nb2/3)O3: PZN). Their crystal structures are complex perovskite; the B-sites of ABO3 are occupied by two chemical elements. Solid solutions of these materials with lead titanate (PbTiO3: PT) exhibit large piezoelectric constants and high mechanical electro couplings (Kuwata et al., 1982). In Zhang et al. (2018), the medical application of PMN-PT is investigated. The use of lead-free piezoelectric materials is preferred owing to the recent increase in environmental considerations. The Restriction of the use of certain Hazardous Substances in electrical and electronic equipment (RoHS) Directive by the European Union came into effect in 2006, and the use of harmful substances such as lead is severely restricted. However, the use of PZT is not prohibited because there is no alternative material for it. However, it is important to develop lead-free alternatives in the near future. Potential candidates include potassium sodium niobate ((K,Na)NbO3: KNN)-based perovskite and sodium bismuth titanate ((Bi1/ 2Na1/2)TiO3: BNT)-based materials (Lin et al., 2005). In Saito et al. (2004), a KNN-based material comparable to PZT is reported.

Sensors based on piezoelectricity General components of piezoelectric sensor Fig. 5A and B shows general piezoelectric sensors and the structure of a piezoelectric sensor, respectively. Electrodes, made of materials such as silver, are attached to opposite faces of the piezoelectric material. When acoustic waves propagate to the sensor, the piezoelectric material is subjected to stress. Then, through the piezoelectric effect, an electric voltage is induced between the electrodes. Thus, by measuring the electric signals, we can detect the acoustic waves. The surface layer of the sensor not only protects the piezoelectric material but also propagates acoustic waves to the piezoelectric material. This is because the acoustic impedance (i.e., the product of the density and magnitude of the acoustic wave velocity) of the surface layer is close to that of the object (Blackstock, 2000). The piezoelectric material is followed by a backing material, which absorbs acoustic waves and changes the vibration frequency of the piezoelectric material, depending on its mass. If we apply voltages to the electrodes of the sensor, stresses are induced through the converse piezoelectric effect; thus, the piezoelectric material can be used as an actuator to generate ultrasonic waves.

Vibration modes Fig. 6 shows the vibration modes that can be used to sense acoustic waves. In Fig. 6A, the polarization of the piezoelectric material is toward direction 3, and compressive stress is being applied in the same direction. In this mode, the d33 components of the piezoelectric constants are primarily utilized, and many piezoelectric sensors are based on this mode. In the mode shown in Fig. 6B,

Fig. 6 Modes for acoustic wave sensing. The polarization is denoted as P, and the stress is T. (A) Thickness mode. (B) Lateral mode. (C) Shear mode. Modified from Turner RC, Fuierer PA, Newnham RE, and Shrout TR (1994) Materials for high temperature acoustic and vibration sensors: A review. Applied Acoustics 41: 299–324.

Physical Sensors: Acoustic Sensors

81

compressive stress is applied in direction 1, which is perpendicular to the polarization. In this case, the d31 constant is utilized. Fig. 6C shows a sensor using a shear mode, in which d15 plays an important role in sensitivity.

Resonance Here, we analytically investigate an excitation of electric variables by periodic stress using a simple model. As shown in Fig. 7, a rod   (length l, width b, thickness h, and density r) is subject to stress on its end faces x1 ¼ 2l . Using angular frequency u and time t, the stress is given by T0eiut, where i is the imaginary unit. For width direction x2 and thickness direction x3, the stresses are free: T2 ¼ T3 ¼ 0:

(7)

The electric field and electric displacement are directed toward the thickness direction: E1 ¼ E2 ¼ 0; E3 s0

(8)

D1 ¼ D2 ¼ 0; D3 s0:

(9)

and

In the above case, the fundamental piezoelectric equations are given by S1 ¼ sE11 T1 þ d31 E3

(10)

D3 ¼ d31 T1 þ 3 T33 E3 :

(11)

and

From Newton’s law, we obtain the following (Rose, 2004): ru2 u1 ¼

vT1 ; vx1

(12)

where u1(x1)eiut is the displacement in the x1 direction. Furthermore, from Maxwell’s equation, rotE ¼ 0, and divD ¼ 0,we obtain vE3 ¼0 vx1

(13)

vD3 ¼ 0: vx3

(14)

and

Since D3 and T1 are independent of x3, we observe that E3 is also independent of x3 from Eq. (11). Thus, E3 can be expressed by the voltage V between the electrodes as V E3 ¼ : h By eliminating T1 from Eq. (12) using Eq. (10), the following is obtained:   v 1 d31 1 v2 u1 ru2 u1 ¼ S  E ; 1 3 ¼ E E E vx1 s11 s11 s11 vx21 where Eq. (13) is used. The general solution of Eq. (16) is expressed by u  u  u1 ¼ A sin x1 þ B cos x1 v v

(15)

(16)

(17)

where A and B are constant, and v is the velocity with which acoustic waves propagate: 1 v ¼ qffiffiffiffiffiffiffiffiffi: rsE11

Fig. 7

Model of a piezoelectric sensor.

(18)

82

Physical Sensors: Acoustic Sensors  By imposing the boundary conditions at the end faces of the rod

 x1 ¼ 2l , the solution of Eq. (16) is determined:

  u  V sE11 T0 þ d31 h sin v x1  : u1 ðx1 Þ ¼ u ul cos v 2v Thus, the velocity at the end face is

  ul   tan   l d31 V 2v _ u1 ; ¼ i T0 þ E rv 2 s11 h

(19)

(20)

where i is the imaginary unit. An electric current I can be obtained from the time differentiation of the surface charge Q, as follows: ! ZZ   2d31 b l ðd31 Þ2 lb I ¼ iuQ ¼ iu D3 dx1 dx2 ¼ E _ u1 þ iu 3 T33  E V: (21) 2 h s11 s11 Further, we consider two different electric conditions: a shorted electrode case and an open electrode case. In the former case, the voltage is zero (V ¼ 0); thus, Eqs. (20) and (21) become   ul   tan l 2v _ u1 (22) ¼ iT0 rv 2 and I¼

  2d31 b_ l u : 1 2 sE11

Hence, the velocity at the end faces and the current are significantly enhanced at the resonant frequencies:   2v 1 u¼ þ n p ðn ¼ 0; 1; 2; .Þ: l 2 At the lowest mode, the half wavelength of elastic waves corresponds to the length of the rod. In the case of open electrodes, the electric current does not flow (I ¼ 0). Hence, from Eqs. (20) and (21), we obtain   ul tan 2v   iT0 l rv _ u1   ¼ ul 2 tan k231 2v   1þ ul 1  k231 2v and V ¼i

  2d31 h _ u1 l ;   usE11 3 T33 1  k231 l 2

(23)

(24)

(25)

(26)

where the electromechanical coupling coefficient (Eq. 3) is used. The velocity at the end faces and the voltage are indefinitely large when     ul 1  k2 ul ¼  2 31 : (27) tan 2v 2v k31 The corresponding angular frequencies can be graphically obtained as the cross point between the two graphs Y ¼ tan X and Y ¼ 

1k231 X: k231

From the above analysis, we can infer that the electric variables are strongly dependent on the angular frequency of the applied external stress. In the above ideal model, electric variables can be infinite because loss is not considered. However, in actual materials, the maximum value is restricted by the loss of the material, as shown in Fig. 8. The steepness of the resonance is characterized by a mechanical quality factor Qm, which is defined by Qm ¼

u0 ; u2  u1

(28)

Physical Sensors: Acoustic Sensors

Fig. 8

83

Typical resonance of piezoelectric sensors.

where u0 is the angular frequency at the peak, and u1 and u2 are the lower and higher angular frequency at p1ffiffi2 of the peak, respectively. The mechanical Q is different for different materials. When designing piezoelectric sensors, it is important to clarify the frequency responses for a given constituent, including the bucking materials and protective layers. We can intuitively understand the behavior of piezoelectric sensors using an equivalent circuit, developed by Mason (1948). However, for a detailed analysis of piezoelectric sensors, FEM software, such as PZFlex and its successor OnScale, are now used by manufacturers.

Macro fiber composite The performance of piezoelectric sensors has improved over time. In 1999, Macro Fiber Composite (MFC) was invented by the NASA Langley Research Center. The MFC is a film-like, lightweight, and flexible sensor. Fig. 9A shows an image of the MFC. Fig. 9B and C shows P1 and P2 types of the MFC structure. Both are composed of PZT fibers positioned parallelly, and each PZT fiber is polarized in the longitudinal direction. In the P1 type, electrodes are attached so that the electric field parallel to the direction of expansion and contraction can be detected. Thus, the P1 type is related to the piezoelectric coefficient d33. In the P2 type, the direction of the electric field is perpendicular to that of the expansion and contraction. Hence, the P2 type is dependent on d31. The MFCs are predominantly sensitive to the longitudinal direction of the PZT fibers owing to the above structures. This directivity of the MFC sensor was experimentally measured in Eaton et al. (2009). Since MFCs can be embedded in structures, their application to SHM is viable. In Di Scalea et al. (2007), the ultrasonic guided waves were sensed using the MFC, which enabled the monitoring of the composite wing skin-to-spar joint in unmanned aerial vehicles. In Matt and Di Scalea (2007), a method to locate the source of the acoustic waves was investigated using directive MFC sensors. In addition, MFC can be used as an actuator to generate ultrasonic guided waves for damage detection (Okabe et al., 2010; Okabe and Nakayama, 2009).

Piezoelectric sensing at high temperatures The need for industrial monitoring at high temperatures has increased; for example, monitoring thermal power plants in operation can reduce downtime costs. In aerospace and aircraft industries, engine monitoring is helpful to improve equipment reliability. Hence, in this section, we discuss the applicability of piezoelectric sensing to high-temperature environments.

(A) (B)

(C)

epoxy

electrode

PZT 14 7 unit: mm

+ +

− −

+ +

+ + − −

+ −

Fig. 9 MFC sensor. (A) Image of MFC (M0714-P2, Smart Material Corporation, Sarasota, FL, United States). (B) P1 type. (C) P2 type. Modified from Sweb Page of Smart Material Corp. https://www.smart-material.com/MFC-product-mainV2.html (Accessed 15 June 2021).

84

Physical Sensors: Acoustic Sensors

Loss of piezoelectricity at high temperatures When the temperature of ferroelectric materials, such as PZT, is increased, their ferroelectricity is lost owing to a transition to the paraelectric phase at a particular temperature. The temperature at which the phase transition occurs is known as the Curie temperature. Thus, we cannot use sensors made of ferroelectric materials beyond the Curie temperature. For example, the Curie temperature of typical PZT is approximately 350  C (the recommended operating temperature is less than 200  C) (Turner et al., 1994). Notably, when a ferroelectric material is heated beyond its Curie temperature and cooled to room temperature, the domain structure appears as shown on the left in Fig. 3. Thus, it cannot be implemented as a sensor, and poling is again required to recover it. Here, we introduce Randau’s phenomenology, which describes the phase transition well. In this model, the free energy F is given as a function of polarization P and temperature T as follows: F¼

a0 b ðT  Tc ÞP 2 þ P 4 ; 4 2

(29)

where a0(> 0), b(> 0), and Tc are constants. The minimum point of the free energy F is a stable state. To determine the stable point, we differentiate Eq. (29): vF ¼ a0 ðT  Tc ÞP þ bP 3 : vP The minimum point is dependent on the temperature T. When T < Tc, the minimum points are rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi a0 ðTc  TÞ; P¼ b

(30)

(31)

and P ¼ 0 is a local maximum point. When T > Tc, P ¼ 0 is a unique minimum point. The graphs of the free energy are shown in Fig. 10. Thus, at a lower temperature (T < Tc), the material is spontaneously polarized (P s 0); however, at a higher temperature (T > Tc), the polarization disappears (P ¼ 0). Furthermore, using the above model, we can obtain the dielectric constant 3 , which is related to a response of the polarization to an electric field E. When an electric field is applied, the free energy changes as follows: F¼

a0 b ðT  Tc ÞP 2 þ P 4  PE; 4 2

(32)

vF ¼ 0, we obtain where the last term is an interaction between the dipole and the electric field. From vP

E ¼ a0 ðT  Tc ÞP þ bP 3 : Using the above equation, the dielectric constant can be calculated as 8 > 1 > >  1 > < a0 ðT  Tc Þ ðT > Tc Þ dE 3  30 ¼ ¼ > dP 1 > > > : 2a0 ðT  Tc Þ ðT < Tc Þ:

(33)

(34)

Hence, the dielectric constant is significantly large around the Curie temperature. In particular, the behavior of the dielectric constants at temperatures higher than the Curie temperature is experimentally known as the Curie Weiss low, which can therefore be explained by Randau’s model.

Fig. 10

Change of free energy at phase transition temperature.

Physical Sensors: Acoustic Sensors

85

The results of the above model (Eqs. 31 and 34) are particularly valid for the phase transition of tri-glycine sulfate (Hoshino et al., 1957). Furthermore, by developing Randau’s model, Devonshire explained the phase transitions of BaTiO3 (Devonshire, 1949).

Piezoelectric materials adaptable for high temperatures To receive ultrasonic waves in high-temperature environments using the piezoelectric effect, we require ferroelectrics that have high Curie temperatures or non-ferroelectric materials that have stable piezoelectricity at high temperatures. Many materials have been investigated for this application (Jiang et al., 2014; Zhang and Yu, 2011). Table 1 shows the properties of these piezoelectric materials. The piezoelectricity is completely lost at the maximum temperature shown in Table 1. In practical implementations, the operating temperature is below the maximum temperature. This is because usage is also restricted by chemical decomposition and the decrease in electric resistivity in addition to the reduction in piezoelectricity at higher temperatures. Known candidate materials for high-temperature sensing include lithium niobate LiNbO3 (LN), gallium orthophosphate GaPO4, aluminum nitride AlN, langasite La3Ga5SiO14 (LGS), and yttrium calcium oxborate YCa4(BO3)2 (YCOB). Among these, LN is ferroelectric, and the other materials are non-ferroelectric. The Curie temperature of LN is considerably higher (1150  C) than that of the other materials, and its electromechanical coupling is also high. At high temperatures, loss of oxygen to the environment may occur. GaPO4 (Krempl et al., 1997) is similar to quartz (SiO2) in crystal structure: half of the Si atoms of quartz are replaced by Ga atoms and the other half by P atoms. Thus, the properties of GaPO4 are similar to those of quartz. However, the a-b phase transition temperature is significantly higher (970  C). AlN can also exhibit piezoelectricity at a high temperature, such as 1150  C (Patel and Nicholson, 1990). AlN is used as a thin film because it is difficult to grow into a large crystal. LGS (Damjanovic, 1998) is characterized by undergoing no phase transition before its melting point (1470  C). The crystal can be grown to a large size; however, its electric resistivity is low. YCOB was first investigated as a laser wavelength conversion element owing to its nonlinear optical effect (Iwai et al., 1997). However, it received more research attention after the discovery of its stable piezoelectric properties and high electric resistivity. To develop a piezoelectric sensing device that is adaptable to high-temperature environments, the heat resistance of electrodes, backing materials, and adhesives between the sensor and the monitoring object should also be considered; for example, instead of metal electrodes, conductive ceramic electrodes may be useful. Using the abovementioned piezoelectric materials, many studies have included experiments based on receiving acoustic waves at high temperatures. In Baba et al. (2010), a transducer based on LN was developed, and ultrasonic waves transmitted through stainless steel were observed at 1000  C. In Kostan et al. (2016), GaPO4 was used to detect a small artificial defect in steel using the pulseecho method at 580  C. Furthermore, Hou et al. (2012) showed that an AlN thin film transducer can be operated at 550  C, and Parks et al. (2013) showed that a YCOB-based transducer can receive ultrasonic waves at 950  C.

Buffer rod method The buffer rod method can be used to receive ultrasonic waves in high-temperature environments (Ihara et al., 2000; Rehman et al., 2001). In this method, a buffer rod mediates between a piezoelectric sensor and a high-temperature substance. One end of the buffer rod is at a high temperature, whereas the other is at a low temperature. Hence, an ordinary piezoelectric sensor can be used in this environment. Ihara et al. developed a buffer rod that comprised of mild steel core and stainless steel cladding; they showed that by adding the cladding, ultrasonic sensing improved because the number of propagating modes was reduced (Ihara et al., 2000). Using the rod, they estimated the cleanness of molten zinc at 600  C by detecting impurities of inclusions.

Table 1

Properties of piezoelectric materials.

Materials

Piezoelectric constant d (pC/N)

a-Quartz Soft PZT PMN-PT LN GaPO4 AlN LGS YCOB

(d11 ¼) 2 (d33 ¼) 417 2500 (d15 ¼) 69 (d11 ¼) 4.5 5.6 6 3–10

Electromechanical coupling k (%) 500 C) ultrasonic transducers: An experimental comparison among three candidate piezoelectric materials. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control 60, 1010–1015. Patel, N.D., Nicholson, P.S., 1990. High frequency, high temperature ultrasonic transducers. NDT International 23, 262–266. Polz, L., Dutz, F.J., Maier, R.R.J., Bartelt, H., Roths, J., 2021. Regenerated fibre Bragg gratings: A critical assessment of more than 20 years of investigations. Optics and Laser Technology 134, 106650. Rehman, A.U., Jen, C.K., Ihara, I., 2001. Ultrasonic probes for high temperature immersion measurements. Measurement Science and Technology 12, 306–312. Rose, J.L., 2004. Ultrasonic Waves in Solid Media, 1st edn. Cambridge University Press, Cambridge. Rosenthal, A., Razansky, D., Ntziachristos, V., 2011. High-sensitivity compact ultrasonic detector based on a pi-phase-shifted fiber Bragg grating. Optics Letters 36, 1833–1835.

96

Physical Sensors: Acoustic Sensors

Saito, Y., Takao, H., Tani, T., et al., 2004. Lead-free piezoceramics. Nature 432, 84–87. Sawaguchi, E., 1953. Ferroelectricity versus antiferroelectricity in the solid solutions of PbZrO3 and PbTiO3. Journal of the Physical Society of Japan 8, 615–629. Shao, L.Y., Canning, J., Wang, T., Cook, K., Tam, H.Y., 2013. Viscosity of silica optical fibres characterized using regenerated gratings. Acta Materialia 61, 6071–6081. Stowe, D.W., Moore, D.R., Priest, R.G., 1982. Polarization fading in fiber interferometric sensors. IEEE Transactions on Microwave Theory and Techniques 30, 1632–1635. Tsuda, H., 2006. Ultrasound and damage detection in CFRP using fiber Bragg grating sensors. Composites Science and Technology 66, 676–683. Tsuda, H., Sato, E., Nakajima, T., et al., 2009. Acoustic emission measurement using a strain-insensitive fiber Bragg grating sensor under varying load conditions. Optics Letters 34, 2942–2944. Turner, R.C., Fuierer, P.A., Newnham, R.E., Shrout, T.R., 1994. Materials for high temperature acoustic and vibration sensors: A review. Applied Acoustics 41, 299–324. Uchino, K., 2015. Glory of piezoelectric perovskites. Science and Technology of Advanced Materials 16, 046001. Urick, R.J., 2013. Principles of Underwater Sound, 3rd edn. Peninsula Publishing, Newport Beach. Valasek, J., 1921. Piezo-electric and allied phenomena in Rochelle salt. Physical Review 17, 475–481. Vul, B.M., 1946. High and ultrahigh dielectric constant materials (in Russian). Electrichestvo 3, 12. Wainer, E., 1946. High titania dielectrics. Transactions of the Electrochemical Society 89, 331. Wang, T., Shao, L.-Y., Canning, J., Cook, K., 2013. Temperature and strain characterization of regenerated gratings. Optics Letters 38, 247–249. Wang, Y., Yuan, H., Liu, X., et al., 2019. A comprehensive study of optical fiber acoustic sensing. IEEE Access 7, 85821–85837. Wee, J., Wells, B., Hackney, D., Bradford, P., Peters, K., 2016. Increasing signal amplitude in fiber Bragg grating detection of Lamb waves using remote bonding. Applied Optics 55, 5564–5569. Wild, G., Hinckley, S., 2008. Acousto-ultrasonic optical fiber sensors: Overview and state-of-the-art. IEEE Sensors Journal 8, 1184–1193. Wu, Q., Okabe, Y., 2012. High-sensitivity ultrasonic phase-shifted fiber Bragg grating balanced sensing system. Optics Express 20, 28353–28362. Wu, Q., Yu, F., Okabe, Y., Kobayashi, S., 2015. Application of a novel optical fiber sensor to detection of acoustic emissions by various damages in CFRP laminates. Smart Materials and Structures 24, 015011. Wu, Q., Okabe, Y., Yu, F., 2018. Ultrasonic structural health monitoring using fiber Bragg grating. Sensors 18, 3395. Wu, Q., Wang, R., Yu, F., Okabe, Y., 2019. Application of an optical fiber sensor for nonlinear ultrasonic evaluation of fatigue crack. IEEE Sensors Journal 19, 4992–4999. Xin, Y., Sun, H., Tian, H., et al., 2016. The use of polyvinylidene fluoride (PVDF) films as sensors for vibration measurement: A brief review. Ferroelectrics 502, 28–42. Yoshino, T., Kurosawa, K., Itoh, K., Ose, T., 1982. Fiber-optic Fabry-Perot interferometer and its sensor applications. IEEE Transactions on Microwave Theory and Techniques 30, 1612–1621. Yu, F., Okabe, Y., 2017. Fiber-optic sensor-based remote acoustic emission measurement in a 1000  C environment. Sensors 17, 2908. Yu, F., Okabe, Y., 2019. Regenerated fiber Bragg grating sensing system for ultrasonic detection in a 900  C environment. Journal of Nondestructive Evaluation, Diagnostics and Prognostics of Engineering Systems 2, 011006. Yu, B., Kim, D.W., Deng, J., Xiao, H., Wang, A., 2003. Fiber Fabry-Perot sensors for detection of partial discharges in power transformers. Applied Optics 42, 3241–3250. Yu, F., Okabe, Y., Wu, Q., Shigeta, N., 2016a. A novel method of identifying damage types in carbon fiber-reinforced plastic cross-ply laminates based on acoustic emission detection using a fiber-optic sensor. Composites Science and Technology 135, 116–122. Yu, F., Okabe, Y., Wu, Q., Shigeta, N., 2016b. Fiber-optic sensor-based remote acoustic emission measurement of composites. Smart Materials and Structures 25, 105033. Yu, F., Saito, O., Okabe, Y., 2020. Laser ultrasonic visualization technique using a fiber-optic Bragg grating ultrasonic sensor with an improved adhesion configuration. Structural Health Monitoring 20, 303–320. Zhang, B., Kahrizi, M., 2007. High-temperature resistance fiber Bragg grating temperature sensor fabrication. IEEE Sensors Journal 7, 586–591. Zhang, S., Yu, F., 2011. Piezoelectric materials for high temperature sensors. Journal of the American Ceramic Society 94, 3153–3170. Zhang, Z., Xu, J., Yang, L., et al., 2018. Design and comparison of PMN-PT single crystals and PZT ceramics based medical phased array ultrasonic transducer. Sensors and Actuators A: Physical 283, 273–281. Zu, H., Wu, H., Wang, Q.M., 2016. High-temperature piezoelectric crystals for acoustic wave sensor applications. IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control 63, 486–505.

Relevant Websites https://ec.europa.eu/environment/topics/waste-and-recycling/rohs-directive_endRoHS. https://onscale.com/dOnScale. https://www.smart-material.com/ContactUsV2.htmldSmart Material Corporation.

Physical Sensors: Magnetic Sensors Marc Christopher Wurz and Maren S. Prediger, Gottfried Wilhelm Leibniz University Hanover, Institute of Micro Production Technology, Garbsen, Germany © 2023 Elsevier Ltd. All rights reserved.

Hall devices Magnetoresistance (MR) AMR and PHE sensors GMR sensors TMR sensors SMR, EMR, CMR Organic semiconductors Magnetostriction and magnetoelasticity Magnetoimpedance Other magnetometers Flux-gate magnetometer SQUID Quantum effect magnetometers Precession magnetometer Optical atomic magnetometers References Relevant Websites

98 99 99 100 102 102 103 103 104 104 104 105 106 106 107 107 110

Glossary Colossal Magnetoresistance (CMR) Occurs in magnetite perovskite materials. Extraordinary Magnetoresistance (EMR) Occurs in semiconductor materials with a metallic inclusion. Giant Magnetoimpedance (GMI) Occurs in soft magnetic, ferromagnetic wires, ribbons and patterned thin film stacks. Giant Magnetoresistance (GMR) Occurs in spin valve systems, amongst others. Hall Effect Phenomenon of a measurable electric potential difference within a metal probe when a magnetic field perpendicular to the plain of the metal probe exists. Magnetoimpedance (MI) A change of intrinsic impedance as a result of an external magnetic field. Magnetometer Instrument used to measure the characteristics of a magnetic field. Magnetoresistance (MR) A change of intrinsic resistance as a result of an external magnetic field. Magnetostriction Change of length in response to an external field. Organic Magnetoresistance (OMAR) Occurs in highly p-conjugated, flexible, conductive, organic materials. Semiconductor Magnetoresistance (SMR) Magnetoresistance in semiconductor materials. Tunnel Magnetoresistance (TMR) Occurs in magnetic tunnel junctions.

Abbreviations AMR Anisotropic Magnetoresistance OPM Optically Pumped Magnetometer SERF Spin exchange relaxation-free magnetometer SQUID Superconducting Quantum Interference Device

Abstract The category magnetic sensor encompasses sensors able to detect a magnetic field and derive characteristics about it, such as the strength of the field, the direction, and the flux. A diversity of sensor types has been developed, based on varying physical phenomena associated with magnetic fields. The chapter provides a general overview of magnetometers based on the Hall

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00054-6

97

98

Physical Sensors: Magnetic Sensors effect, magnetoresistance, magnetostriction, magnetoimpedance, as well as quantum phenomena. It briefly describes them per their working principle and comments on common materials and applications.

The category magnetic sensor encompasses systems able to detect a magnetic field and derive characteristics about it, such as the strength of the field, the direction, and the flux (Popovic et al., 1996). The term magnetometer denotes devices in general designed to measure these characteristics. On the global market, magnetic sensors are predominantly used in the fields of automotive, consumer electronics, industrial applications and aerospace. Non-destructive testing and medical research and development, often within the market share category others, have raised their use of magnetic devices in recent years as well. The total market value of magnetic sensors surpassed 2 billion US$ in 2018 and was estimated at 4.3 billion US$ in 2020. By 2025, with an estimated compound annual growth rate between 7% and 8%, the market value is expected to increase up to 6.2 billion US$. The dominating technologies are systems based on the Hall Effect, followed by magnetoresistive systems, commonly referred to as xMR devices. Fluxgate sensors and SQUID instruments hold a margin of the market, while other technologies tend to be listed cumulatively as others (Dixon, 2019; Palsule et al., 2019; Wadhwani et al., 2019, 2020). A complete magnetic sensor microsystem does not only contain the magnetic sensor unit itself, but also biasing components, signal-processing parts, circuitry and optionally coils, permanent magnets, and magnetic concentrators (Popovic et al., 1996; Calkins et al., 2007). A diversity of sensor types has been developed, based on varying physical phenomena associated with magnetic fields. Following, these sensor types are briefly described per their working principle and common applications. Please note, that the list is not exhaustive and the author selected the principles based on personal judgment of importance.

Hall devices As mentioned, the majority of commercially available magnetic sensors are Hall devices, which are based on the Hall Effect. The Hall Effect itself describes the phenomenon of a measurable electric potential difference within a metal probe when a magnetic field perpendicular to the plain of the probe exists. The basis of this effect is the Lorentz force; a force acting on moving charge carriers in an electromagnetic field. It is important to point out, that this effect is an electrical manifestation of a magnetic field acting on conducting electrons in conducting materials (Akouala et al., 2019). When a current is applied in-plane to the probe, there is always a magnetic field perpendicular to the plane associated with said current. Any magnetic field exerts the Lorentz force on the current charge carriers in the metal probe, as shown in Fig. 1 (left). Therefore, solely by running current through the probe, the magnetic field deflects charge carriers laterally, leading to an area more negatively charged within the metal (and more positively charged due to depletion) and subsequently to a potential gradient across the plane of the probe. When the metal probe is electrically contacted in plane and perpendicular to the current, the gradient can be measured as a voltage. Any additional, external magnetic field that is also oriented perpendicular to the probe plane will increase the gradient, which makes the external field measureable as a function of the voltage and the known current. Additionally, the voltage polarity depends on the direction of the magnetic field. Goel et al. (2020) list the p-type semiconductors InSb, InP, GaAs as well as graphene as the most suitable materials for Hall sensors. These types of sensors, however, are temperature dependent and may show an offset-voltage due to the magnetization from the current as well as a voltage drift. Applications for magnetic field sensing with Hall probes range from speed and angle sensing to image stabilization in smartphones. Commercially available systems are usually Hall sensors with integrated circuitry (Hall IC) to include components such as voltage regulators and amplifiers. Hall ICs are commonly designed to function as a latch or switch (uni-, bi-, or omnipolar) and can be operated with either of two settings: constant current drive and constant voltage drive. The authors recommend Goel et al.’s review (2020) on operating mode and the web references for graphical representations.

Fig. 1 (Left) When a current I is applied in plane of the Hall material, it creates a magnetic field in the direction M. The associated Lorentz force FL deflects the charge carriers, leading to a potential UI across the material. Filled charge carriers represent electrons, lighter circles represent electron depletion. (Right) Contacting of a Hall probe.

Physical Sensors: Magnetic Sensors

99

Magnetoresistance (MR) Semiconductor, anisotropic, giant, tunnel, colossal and extraordinary magnetoresistive sensors all rely on the principle that the intrinsic electrical resistance changes due to an external magnetic field. Zheng et al. (2019), report on various applications of MR sensors: detection of the earth’s magnetic field for navigation and transportation purposes, non-destructive monitoring of power-grids, detection of ultra-low magnetic fields in biomedical applications from bio-functionalized nanoparticles or nanostructures and human signals, and integration in wearable devices and autonomic vehicles. Generally, Zheng et al. (2019) cluster the applications into field strength categories of mT range, mT to nT range, and the range of nT and below. A prominent application is the integration of giant magnetoresistive sensors in read heads for data storage through magnetic recording and recently, the extraordinary magnetoresistance has been studied for ultrahigh-density magnetic recording (Hewett and Kusmartsev, 2010). The effect of magnetoresistance is commonly denoted as the ratio of the change in resistance in a field (DR(B)) to the resistance in zero magnetic field (R0). The term of change in resistance therein is calculated as the difference between the resistance in a field and the resistance in zero field (DR(B)]R(B)–R0) (Pippard, 1989). The sign of magnetoresistance can therefore be negative or positive, depending on whether the field increases or decreases the material’s resistance compared to zero field (Solin et al., 2000). Magnetoresistance has been observed in various materials with a wide range of magnitudes in DR(B). Ali et al. (2014) states that, in principle, negative values are found in magnetic materials and positive values in metals, semimetals, and semiconductors. Organic semiconductors exert magnetoresistance of both signs, too (Bloom et al., 2007; Hu and Wu, 2007; Joshi et al., 2018). Controlling and tuning the extent of magnetoresistance in materials is currently a focus of research, through material optimization and voltage control of magnetic layers, for example (Nichterwitz et al., 2020). This way, a switching between negative and positive magnetoresistance within the same material is possible. The abbreviations AMR, GMR, TMR, EMR and so on tend to, on one side, describe the magnitude of the MR effect, but also, on the other side, differentiate between the materials used and the underlying physical cause of magnetoresistance. The following attempts to provide a general understanding of this vast field.

AMR and PHE sensors These two sensors are technically two sides of the same coin, as their measurement principle is caused by the scattering of conducting charge carriers (Groenland et al., 1992) at the atomic spin-orbitals of ferromagnetic materials depending on the angular in-plane relationship between the material’s magnetization and the electrical current direction (Burkov, 2017; Li et al., 2018a). Ferromagnetic materials align their magnetization orientation with the orientation of an external magnetic field. Two states are commonly differentiated: firstly, the magnetic field and therefore the internal magnetization being perpendicular to the electric current and secondly, the magnetization and current being parallel (Li et al., 2018a). In the former case, using a simplified approach depicted in Fig. 2A, the atomic spin orbitals of the metal atoms are aligned with the direction of the current, yielding a minimal scattering of charge carriers. Accordingly, when the magnetization is parallel to the current, the spin orbitals are perpendicular to the current and scattering of charge carriers is maximized (Fig. 2B). Solely by definition, this dependence of electrical resistivity on the angle between magnetization and applied current is called anisotropic magnetoresistance. Depending on in which direction the measurement is taken in relation to the current, the longitudinal resistivity and the transverse resistivity are distinguished. In literature, however, AMR sensors commonly only refer to measuring the change in resistivity longitudinally. The term planar Hall effect (PHE) denotes the change in transverse direction of the current (Groenland et al., 1992). AMR and PHE therefore are based on the same physical effect (Schuhl et al., 1995; Akouala et al., 2019). This is illustrated in Fig. 2C Zheng et al. (2019), divide AMR sensors in parallel and perpendicular AMR sensors, not making use of the misnomer PHE. For the ease of differentiation, parallel AMR herein are called AMR and perpendicular PHE. For AMR sensors, the change in resistivity is measured as the change in voltage in the direction of the current. For a mathematical derivation, please refer to Kokado and Tsunod (2013). The most prominent material for the purpose of AMR sensors is permalloy in the form of thin films. The alloy’s advantage is negligible magnetostriction (refer to the section magnetostriction) and AMR ratios of up to 3%. To accomplish high AMR ratios, it is important that the magnetization in the sensor layer is as uniform as possible and,

Fig. 2 (A) Spin orbital orientation for the AMR effect with the magnetic field orientation in plane parallel (topdhigh resistance) and perpendicular (bottomdlow resistance) to the current; (B) Schematic depiction of the difference between PHE (top) and AMR (bottom): direction of resistance measurement.

100

Physical Sensors: Magnetic Sensors

without an external magnetic field, parallel to the current direction. In an ideal scenario, the material is a single magnetic domain. The resistivity is then at a maximum and an external magnetic field causes a reduction of the resistivity. Sensor designs take into account the magnetic domain size by using thin films and meandering geometries, which introduce a shape anisotropy and increase the longitudinal domain component in relation to the transverse domain component (compare Fig. 3). Additionally, meandering geometries raise the basal resistivity of the sensor and the effective length, thereby amplifying the signal. With the addition of socalled barber poles (stripes of conductive material 45 degrees across the sensor material) the sensor signal is linearized (Groenland et al., 1992; Ripka, 1994; Ripka and Arafat, 2019; Demirci, 2020). Daisy chain designs, where the meander turns (the areas of transverse domains) are substituted with non-magnetic conductive material such as copper, have been investigated as a way to eliminate the signal component from transverse magnetic domains. Commercial sensor systems feature Wheatstone-Bridge layouts. Besides permalloy, anisotropic magnetoresistance can be found in NiCr and NiFeCr alloys. Topological Weyl semimetals show a negative longitudinal magnetoresistance and Dirac semimetal ribbons of Cd3As2 can reach AMR ratios of 68% (Li et al., 2018a). AMR sensors are commercially fabricated on a silicon substrate; other substrate materials found in literature include flexible substrates such as polyimide or the thermoplastic compound polyether etherketone (PEEK) and insulating glass (Jogschies et al., 2014; Krischik and Rissing, 2015; Rittinger et al., 2015; Prediger et al., 2019a,b; Wittek, 2019; Aue, 2020). The PHE has been studied for low-field magnetic sensor applications due to its high sensitivity regarding the magnetization direction and for its use in compasses and magnetic biosensors (Akouala et al., 2019). Ejsing et al. (2004), and Damsgaard et al. (2008), for example, realized micro- and nanobead detectors. It is also a powerful tool in studies of exchange coupling occurring between magnetic thin films and in the characterization of undoped magnetic oxide thin films. For PHE sensors, prominent geometric designs are the resistor element, the cross-shaped sensors, the planar Hall effect bridge (PHEB) and the meander version of it (mPHEB) (Henriksen et al., 2010). The planar Hall effect can not only be observed in metallic ferromagnets such as nickel, iron and cobalt, but also in ferromagnetic semiconductors and the oxide ceramic lanthanum strontium manganite (LSMO) (Bason et al., 2004). In materials such as epitaxial GaAs or MnAs thin films, the extent of PHE can be up to four magnitudes higher, so that the effect was coined the giant planar Hall effect (GPHE) (Tang et al., 2003).

GMR sensors The macroscopic effect of resistance switching in giant magnetoresistance (GMR) devices stems from the interaction of two magnetic layers, the relationship of their magnetic moments and the spin polarization of current carriers. In the simplest setup, a conductive region separates two ferromagnetic regions as seen in Fig. 4. Two situations are described: in the first, the magnetic moments of the ferromagnets align; in the second, the magnetic moments are antiparallel. When considering the majority of the current carriers to have a spin polarization in accordance with the left ferromagnet, the first situation results in low resistance and the second in high resistance. The current carriers in the second situation are scattered at the interface from the conductive region to the right ferromagnet, therefore hindering conductance between the two ferromagnets and causing a high increase in magnetoresistance. An external magnetic field can easily change the orientation of a ferromagnet. When assuming that one of the ferromagnet’s magnetic

Fig. 3 Sensor Designs: (A) Meander, permalloy on glass substrate, scale bar equals 540 mm; (B) example of a Wheatstone bridge layout, inset shows the technical configuration; scale bar equals 100 mm; (C) barber pole structure: Au barber poles on a Cr adhesion layer deposited in 45 angle on Pt conductor strip, scale bar equals 20 mm; kindly supplied by E. Demirci, Department of Physics, Gebze Technical University; previously published in Demirci (2020), Figure 1b; (D) technical drawing of AMR sensor in daisy chain layout, front side (left) features permalloy strips connected through copper VIAs to the backside (right) interconnects and contact pads; developed at the Institute of Microproduction Technology, scale bar equals 900 mm.

Physical Sensors: Magnetic Sensors

101

Fig. 4 Spin dependent crossing between the two ferromagnets across a conductive region with the orientation of the two ferromagnetic in parallel (left) and antiparallel (right). The lower scheme shows the distribution of spin population on the energy scale close to the Fermi level and the transitions. For the antiparallel case, the spin transition is not possible. Modified from Prinz GA (1998) Magnetoelectronics. Science 282: 1660–1663.

moment is fixed and the other one free to move with an external field, the resistance between the two serves as a measure of the external field (Prinz, 1998; Bürgler, 2010). For a more in-depth discussion of the mechanisms underlying giant magnetoresistance, the authors recommend Bürgler (2010) and Prinz (1998). Since its discovery in 1988, GMR devices were laid out as anti-ferromagnetically coupled multilayer stacks and double coercivity ones. The spin-valve configuration, however, is the current state of the art. For a short historic review, please refer to Baraduc et al. (2013). The basic structure of a spin-valve is a sandwich of a non-magnetic but electrically conductive layer between two magnetic layers. The magnetic orientation of one layer is purposefully pinned with an additional anti-ferromagnet (Fig. 5A). From the basic structure, additional layers can be introduced into the stack to increase the magnetoresistive ratio. For example, in order to improve the magnetic stability of the pinning anti-ferromagnet, a combination of layers creates a so-called synthetic antiferromagnetic pinned layer together with the to-be-pinned ferromagnetic layer (Fig. 5B). One can also improve the MR ratio by increasing the exchange stiffness on the interfaces between the antiferromagnetic layers and the non-magnetic electrically-conductive layer (Reig et al., 2013). When using a GMR layer stack, the layers can be electrically connected in two ways: current in plane (CIP) and current perpendicular to plane (CPP). CIP is historically employed, although the giant magnetoresistive effect is only a secondary effect, since the current travels parallel along the layers. In order to build a model that explains experimental results, several terms need to be considered, including the spin dependent reflection and the transmission coefficients at the interfaces, which makes a theoretical description complex. When the stack is contacted perpendicular to the plane of the layers, the current travels across the layers. A simple model of the current in series in a two-channel resistor describes empirical CPP results well (Bürgler, 2010; Reig et al., 2013). In

Fig. 5 (A) Basic spin valve thin film stack consisting of a protective layer, the free layer, which changes magnetization with the external field, the conductive spacer layer, the pinned layer with a magnetization direction set by the pinning layer; (B) spin valve stack with synthetic anti-ferromagnet consisting of Co and Ru layers and an additional boundary layer between the free layer and the spacer; materials and thicknesses are examples only.

102

Physical Sensors: Magnetic Sensors

addition to the spin-valve, hybrid nanostructures of semiconductor and magnetic materials exhibit giant magnetoresistance as well. The materials are termed magnetically confined semiconductor heterostructures (MCSH) (Lu et al., 2016). GMR sensors have been incorporated in automotive applications (angle, speed, and position measurement) as well as in GPS systems. They are used in nondestructive evaluation processes, biomedicine, imaging and scanning microscopy, and the measurement of electrical currents at the integrated circuit level (Reig and Cubells-Beltran, 2013).

TMR sensors Like the Giant Magnetoresistance, the Tunnel Magnetoresistance (TMR) is a spintronic phenomenon observed in so-called magnetic tunnel junctions (MTJs). The basic system differs from a GMR system in the use of an insulator (or semiconductor) instead of a conductor as the spacer material between two magnetic regions. The material builds a barrier that can be crossed by current carriers through spin-polarized quantum-mechanical tunneling alone. An electron can tunnel through the barrier only when a state with the same spin-polarization is unoccupied on the other side. The contribution to a tunneling current is higher when the magnetization on both sides of the spacer are aligned, thereby reducing the resistance across the insulator (Bürgler, 2010). Common spacer materials are oxides of aluminum, titanium, and magnesium. Literature divides MTJs in incoherent and coherent ones, where the distinction is based on the lattice formation and matching of the magnetic and insulator layers (Knudde et al., 2017). During fabrication, temperature annealing of the stacks favors the alignment of crystal lattices in adjacent materials (Knudde et al., 2017), which increases TMR values. Parkin et al. (2004), report 200% TMR values at room temperature with CoFe ferromagnets and magnesium oxide as the tunnel barrier; up to thousands of percent of MR were observed in material combinations of body centered cubic ferromagnetic layers and magnesium oxide as insulator (Reig et al., 2013). With the push for more flexible substrates, Barraud et al. (2010) fabricated TMR sensors on the organic flexible substrate poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) (PEDOT-PSS), reaching a TMR ratio of 12.5%. In application, hundreds of TMR stacks are employed for increased robustness (Vidal et al., 2017) and the technology is under study for areas similar to GMR sensors (Leitao et al., 2015; Weitensfelder et al., 2018a). Most recently, a new design of GMR and TMR layer stacks with a circular free layer has been researched. The circular layer favors a closed, just as circular vortex magnetization where the vortex nucleation itself can turn out of the layer plane to minimize energies as seen in Fig. 6. This design is nearly hysteresis free, a crucial advantage for any application (Pigeau, 2012; Wurft et al., 2017; Weitensfelder et al., 2018b; Wurft, 2018; Schüthe et al., 2019).

SMR, EMR, CMR High-mobility semiconductors can exhibit magnetoresistance as well (SMR), such as linear magnetoresistance without saturation in composite materials of disordered MnAs-GaAs (Johnson et al., 2010) or anisotropic magnetoresistance in antiferromagnetic semiconductors like Sr2IrO4 (Fina et al., 2014). Mell and Stuke (1970) and Popovic et al. (1996) attribute MR in semiconductors to the so-called physical magnetoresistive effect. In comparison to the geometric magnetoresistive effect, where the resistance changes because of a turn of magnetization, the physical magnetoresistive effect causes an orbital deflection of internal charge carriers due to the magnetic part of the Lorentz force. Extraordinary magnetoresistance (EMR) of more than 1 million percent can be observed in semiconductors with a metallic inclusion, such as in the combination of InSb and Au. The mechanism for this magnetoresistance is a change in current distribution through the two material areas. Studies have used disc-like designs, with a circular metal inclusion in the middle, and designs with branched inclusion geometries, achieving a higher effect with the latter (Hewett and Kusmartsev, 2010; Sun and Kosel, 2013). A non-magnetic behavior, such as caused by the phase transition between paramagnetic and ferromagnetic, results in colossal magnetoresistance (CMR) in 3d transition metal oxides (Ramirez, 1997). Manganite perovskites of various compositions demonstrate MR ratios in the thousand and hundred thousand percent (Jin et al., 1994; Solin et al., 2000). Dagotto (2003) provides a detailed record on CMR and Balevicius et al. (2019), describe the design of a CMR probe for a handheld magnetic field sensor.

Fig. 6 Vortex state simulation on a circular free layer; left: the external magnetization is positive; middle: no external magnetization; right: the external magnetization is negative. Reprinted and modified from Pigeau B (2012) Magnetic Vortex Dynamics Nanostructures, p. IX, Université Paris Sud: Paris.

Physical Sensors: Magnetic Sensors

103

Organic semiconductors In the last two decades, a new field of research with respect to magnetoresistance emerged. Here, organic materials are studied, on one hand, for their use as spacer layers between electrodes in a spin-valve setup and, on the other hand, for their intrinsic change of properties associated with an applied magnetic field; amongst electroluminescence and photocurrent, their change in resistance. This lead to coining the term organic magnetoresistance (OMAR). For spin-valve systems, the organic layer serves as a spacer similar to GMR and TMR devices. Materials such as conducting polyaniline-carbon nanotubes-composites, conducting PEDOT-PSS, tris-8hydroxyquinoline aluminum (Alq3), polyphenylene vinylene (PPV), sexithiophene (T6), copper phthalocyanine (CuPc), fullerene, multilayer graphene and carbon nanotubes in various fashions (multiwalled, bending) are commonly used. The important feature is high p-conjugation in the material, to facilitate a certain degree of charge carrier mobility. In spin-valve devices, spin-polarized charges are injected into the organic material from ferromagnetic electrodes, such as LSMO, NiFe, Ni, Fe, and Co. To reduce the energy mismatch between the electrodes and the organic layer spacer, thin buffer films of metallic oxides, metallic fluorides, or self-assembled monolayers can be integrated in the stack (Bloom et al., 2007; Joshi et al., 2018; Guo et al., 2019; Tanty et al., 2019; Tian and Xie, 2019). The fundamental physical concepts underlying organic magnetoresistance are, however, much more complicated than with inorganic materials (Tanty et al., 2019). The so-called hyperfine magnetic fields from an atomic nucleus influence the spin rotation of charge carriers in the material. Simplified, when an electron with a set spin enters the material, it will hop from one site to another and, with different fields at different atomic nuclei, experience spin relaxation on its way through the material. Because the spin coupling is weak in organic materials, the spin relaxation time is longer than in inorganic materials, which is beneficial for application development. An external magnetic field is easily bigger than hyperfine fields and uniform. Therefore, while carriers are hopping through organic materials with an applied external field, spin relaxation is suppressed and spin polarization transport prolonged. The charge carriers are then, depending on the spin polarization relationship between the two electrodes, scattered at the second interface or can enter the second electrode. This equals the GMR system. Tunneling through the organic material leads to organic tunnel junctions. For reviews on the models and physical background on organic spintronics, the authors suggest the following publications: Wagemans and Koopmans (2011), Tian and Xie (2019) and Shumilin (2020). The use of organic materials for magnetoresistive sensing is especially interesting for the development of flexible electronics and biomedical devices, due to the mechanical flexibility, the biocompatibility and the low-cost production options of the materials. Lastly, for current state-of-the-art research on magnetoresistive sensors, please refer to the section “Sensor Development: Magnetoresistive Sensors” or the excellent review on spintronic sensors from Freitas et al. (2016).

Magnetostriction and magnetoelasticity Based on observations made in the 19th century, the Joule effect, which describes a longitudinal change in length of a material in response to an applied magnetic field, is the foundation for magnetostrictive sensors and actuators. Actuators, however, are beyond the scope of this text. So is the use of the reverse effect, Villari effect, where a stress induces a change of magnetization, which lays ground for force and strain sensors. When a helical magnetic field causes a twisting of the material, the phenomenon is referred to as the Wiedemann effect, also a magnetostrictive effect. The inverse effect (Matteucci effect) can be used to measure torque with helically magnetized amorphous wires, which will not be covered herein. All mentioned effects are summarized under the term magnetoelasticity, which describes the coupling of magnetic characteristic with the elastic ones (Lee, 1955; Calkins et al., 2007; Ekreem et al., 2007; Ren et al., 2019). The elongation in the Joule effect is a result of the realignment of individual magnetic moments and the movement of domain walls under an influence of an ambient magnetic field. Gomonay and Loktev (2002), describe the process that links to increasing aligned domains at the expense of decreasing unaligned domains. A rotation of the magnetic moment is energetically favorable to a complete flop of the spin. Such a rotation affects the orientation of atomic orbitals of the higher order, since they rotate as well. The interatomic distance between the materials atoms therefore increases or decreases locally, depending on the rotational direction and the orientation at zero field. Local changes compound for the entire magnetostrictive structure. The ratio of the longitudinal change and the length at zero field, DL/L, describes the magnetostriction (l). For positive magnetostriction, as found in iron, the material elongates, while it shrinks for negative magnetostriction, for example in nickel (Ekreem et al., 2007; Liu et al., 2012; Vincent et al., 2020). The binary FeGa alloys Galfenol demonstrate moderate magnetostrictive behavior with 350 ppm; meanwhile, Terfenol-D (Tb0.3DY0.7FE1.92) is known as the giant magnetostrictive alloy with approximately 2000 ppm (Atulasimha and Flatau, 2011). Materials for thin-film systems are commonly the alloys FeGaB and FeCoSiB. Magnetostrictive magnetometers are predominantly used for ac field detection (Tu et al., 2019). Thin films of magnetostrictive materials on a cantilever, for example, yield a magnetometer that deduces the magnetic field as a function of voltage directly. Alternatively, the probe material can be pre-strained and the change of permeability as a function of resonance frequency due to an external field is acquired. For indirect detection of the elongation due to an ambient magnetic field, the magnetostrictive material needs to be combined with something that produces a signal due to the elongation. One popular mechanism is to coat optical fibers with a magnetostrictive material or to couple a bulk material with an optical fiber. Both leads to the modulation of the optical path length. Another method is the use of laser interferometry, which falls under the category of a dilatometer, an instrument for elongation measurements. The deposition of strain gauges on the crystalline or amorphous magnetostrictive material and the

104

Physical Sensors: Magnetic Sensors

measurement of the strain gauge’s electrical resistance describes another indirect method. Sensitivities can reach the pico Tesla region, depending on the frequency. The use for biomedical applications has been demonstrated for cardiologic measurements with a cantilever system and magnetic nanoparticle mapping was possible with thin-film magnetometers (Calkins et al., 2007; Ekreem et al., 2007; Tu et al., 2019). Magnetostrictive layers are sometimes incorporated in GMR spin valves, which turns the system into a strain sensor through the change in magnetoresistance at a constant field (Liu et al., 2016).

Magnetoimpedance Within the last 20 years, systems based on an effect termed giant magnetoimpedance (GMI) were studied and developed for their use in sensing direct-current magnetic fields (Knobel and Pirota, 2002; Alam et al., 2020; Janutka and Brzuszek, 2020), especially for biomedical applications with low-frequency fields and pico-Tesla resolution (Zhukov et al., 2015). Li et al. (2015), suggested the use of magnetoimpedance stripes for spatial position sensing of a moving body, while making use of the varying influence of the earth’s magnetic field on the sensing stripe during movement. The basis of the effect is a quite complex topic. In general, the impedance of GMI materials varies with an external magnetic field. Simplified explanations name the strength of the skin effect as the underlying physical effect; therefore, GMI is only observed when the materials are operated with alternating current. As already described above, when a current flows through a conductor, a magnetic field is generated. The magnetic field in turn induces eddy currents in the conductor. The eddy currents cancel the alternating current in the center of the conductor and increase the current at the outer shell, forming a skin of high current density. The skin depth, which is inversely proportional to the driving current frequency (Li et al., 2015), is also termed the penetration depth of the ac current. A material’s impedance depends strongly on the penetration depth, which can be modulated by an external magnetic field (assuming the ac current and frequency is constant), thereby rendering impedance useful for the detection of external magnetic fields. The skin effect limits the possible miniaturization of GMI systems, when the skin depth exceeds the dimensions of the material in question (Janutka and Brzuszek, 2020). The detailed physical description of magnetoimpedance, however, is more complex. Depending on the frequency of the driving current, the general magnetic behavior influencing the complex impedance contributes differently to the total magnetoimpedance. Knobel and Pirota (2002), divide the frequency range into three regimes: From 1 to 10 kHz, the magneto-inductive effect has the largest contribution. Up to a few MHz, the variation of the skin effect through the external field is the driving phenomenon. From MHz to GHz, mainly ferromagnetic resonance causes magnetoimpedance. Zhukov et al. (2015), add an additional fourth regime, dividing Knobel and Pirota’s second regime in two parts. This is a result of the differentiation, which component contributes to the skin effect. In the lower region (10 kHZ–10 MHz) domain wall motion and magnetization rotation cause the skin effect. In the upper region (10 MHz–1000 MHz) the magnetization rotation outweighs domain wall motion due to damping. Analogous to magnetoresistance, the effect of magnetoimpedance is defined as the ratio of the change of impedance DZ and the basal impedance Z (Zhukov et al., 2020). GMI has been studied in Co-rich and Fe-rich soft ferromagnetic materials, amorphous and nanocrystalline, fabricated into magnetic microwires, which can also be glass-coated, or patterned into ribbons (Yang et al., 2019; Alam et al., 2020; Zhukov et al., 2020). Patterning of these structures, such as into a meander shape as shown in Fig. 7A and B, is a several-steps process that includes the chemical etching and polishing after manufacturing of the base material (Yang et al., 2019). The integration of microwires and the necessary coils onto ASIC systems is possible, yet elaborate, with MEMS technologies. An exemplary process is given in Fig. 7C. Thin film stacks of soft ferromagnetic layers and nonmagnetic conductive layers have been employed as well and also realized on flexible substrates (Li et al., 2015; Zhukov et al., 2015). While, usually, materials are thermally annealed to enhance domain structure, this type of annealing typically goes hand in hand with a magnetic hardening of the material. This is counter-productive for magnetoimpedance. Stress annealing and Joule Heating, however, enhance magnetic softness (Zhukova et al., 2018; Zhukov et al., 2020). An excellent review on manufacturing techniques and materials of GMI was provided by Zhukov et al. (2015).

Other magnetometers This section covers specific working principles of magnetometers, which are less commonly used in automotive, industrial and consumer electronic settings, but more commonly in geomagnetic measurements, biomedical applications, and research on fundamental magnetic behavior and symmetry. Especially the later on described quantum magnetometers are suited for biomagnetic measurements and symmetry studies, where sensitivities in the regions of femto, pico and even atto Tesla are required.

Flux-gate magnetometer Amongst the so-far discussed sensor principles, the fluxgate is one of the older principles, originating in the 1930s. It’s source is the change of permeability of ferromagnetic materials in a periodically changing magnetic field. An excitation coil commonly establishes such a field along the axis of the core. The inducing current is switched periodically, so that the excitation field switches direction and changes/shuffles the core from one point of saturation through zero field to saturation again. The time it takes between the

Physical Sensors: Magnetic Sensors

105

Fig. 7 Patterned meander of Co75Fe10Ni2Si8B5 with three (A) and six (B) turns; (C) exemplary process flow for the integration of microwires and coils on ASICs. (A) and (B) Modified from Yang Z et al. (2019) Magnetoimpedance effect in the ribbon-based patterned soft ferromagnetic meandershaped elements for sensor application. Sensors 19(2468).; (C) Modified from Honkura Y and Honkura S (2020) The development of a micro-coil-onASIC type GSR sensor driven by GHz pulse current. Journal of Magnetism and Magnetic Materials 513: 1–13.

points of saturation is defined as the residence time, Tþ and T for each direction (positive to negative, negative to the next positive). The changing magnetic flux in the core induces a voltage in a second coil, which is symmetrical with odd harmonics when no external field is present. Any ambient field adds to the flux in one direction, due to the saturation state with no increase in permeability, which skews the voltage induced in that half-period compared to the other direction. This results, on one hand, in even voltage harmonics and in a difference in residence time between the saturation points, on the other hand. Determination of the former has been, initially and still is, used as the readout signal for fluxgates, the latter, however, is used in devices termed RTD fluxgates (Primdahl, 1979; Andò et al., 2006; Djamal et al., 2011; Nikitin et al., 2013; Tumanski, 2013; Pang et al., 2017). For a detailed introduction, the reader is referred to Ripka (1992). Fluxgate sensors are designed in a variety of ways, which include: circular rod cores, one or two, with both being inside a pick-up coil or each having an individual pick-up coil; ring core structures, with excitation coils wound around the halves of the ring core in opposite direction, which increases sensitivity and reduces power consumption and noise; single cores with both coils around it; thin-film based printed circuit board setups or racetrack structures. General requirements for core materials are soft-magnetic properties, ideally without magnetostriction. The literature reports one the use of permalloy strips wound to rings, ferrites or FeSiB amorphous wires as rod cores, flexible cores of CoFeSiB alloys, amorphous ribbons, and metallic glasses (Ripka, 1992; Andò et al., 2010; Butvin et al., 2012; Trigona et al., 2018). Technology improvements are concerned with the reduction of size, costs and the compatibility with MEMS technology. Due to its nature, the fluxgate is suited to sense weak magnetic fields with low frequencies, similar to the earth’s magnetic field and its ruggedness compared with other devices has landed it seats on space instruments. Fluxgate magnetometers can reach sensitivities in the pico Tesla regime. Research efforts also focus on the detection of metal concentration in brain regions for degenerative disease diagnosis and other biomagnetic detection (Lühr et al., 1985; Díaz-Michelena, 2009; Djamal et al., 2011; Trigona et al., 2018, 2020).

SQUID For years, superconducting quantum interference devices (SQUID) have been the most sensitive systems to measure an ambient magnetic field (Gemmel et al., 2010). Most devices make use of two phenomena: the Meissner Effect and the Josephson Effect. The first is based on magnetic flux quantization in a closed superconducting loop, which, strongly simplified, means that only magnetic flux characterized by being an integral multiple of a flux quantum can penetrate the superconductor, otherwise characterized magnetic flux is effectively expelled from the superconductor. Materials for SQUID devices are operated around the transition temperature TC below which their behavior switches to a superconducting state, so that at the switch, the expelling of the magnetic flux causes a change in voltage proportional to the ambient magnetic flux. The Josephson Effect relates to tunneling of electrons between two superconducting regions across a non-superconducting region. Tunneling is possible up to a critical current density above which the non-superconducting region becomes ordinarily resistive, thereby reducing the superconducting current between the superconductive material. Such regions are denoted as weak links in the loop. Literature shows a multitude of designs for weak

106

Physical Sensors: Magnetic Sensors

links, including point contacts, micro bridges (Dayem bridges), grain boundaries in the form of bicrystal systems or over a step edge, and thin-film tunnel junctions (Josephson Junctions, JJ) with either an insulating or normal metal barrier as weak link (Stuart, 1972; Clarke, 1989; Fagaly, 2006; Zhang et al., 2018). SQUID systems operate with direct current and alternating current in radio frequency, although the direct current models show a better sensitivity. Since superconductivity requires very low temperatures, cooling is necessary. The early materials, such as indium, tin, tantalum, lead, niobium, and niobium-tin, require cooling with liquid helium to reach TC, an expensive and limited resource. Research in the recent years has focused on developing SQUID devices with higher TC materials such as the ceramic yttrium oxide YBCO (YBa2Cu3O7-d) in order to reduce costs by enabling the use of liquid nitrogen as coolant. Two prominent application areas are nondestructive testing and medical biomagnetic measurements, for which high-TC materials are especially interesting (Vaidhyanathan et al., 2011; Storm et al., 2017; Li et al., 2018b; Trabaldo et al., 2020). The interested reader is referred to Fagaly’s (2006) excellent review on the SQUID instrumentation and applications.

Quantum effect magnetometers The only class of magnetometers that can match or top the sensitivity of SQUID devices are quantum effect magnetometers. Similar to SQUID, the main application areas are the earth’s magnetic field and biomagnetic applications, including the detection of the magnetic field from, for example, heart or brain in magnetocardiography and magnetoencephalography, respectively, and the detection of NMR and MRI signals. This magnetometer class spans a variety of systems and is accompanied by a, sometimes confusing, terminology. From atomic magnetometers to Overhauser magnetometers, optical magnetometry and optically pumped magnetometers, spin-exchange relaxation free magnetometers to nuclear and proton precession magnetometers. The following aims to provide a simplified overview of the relations of these terms and systems. The basis for quantum effect magnetometers is the gyromagnetic effect. An atom, as well as the nucleus within an atom and the electrons, can be characterized by an individual angular momentum and magnetic momentum. When the internal magnetic momentum increases, the angular momentum increases and vice versa, defined by the gyromagnetic ratio of the magnetic to the angular momentum. A change in the direction of the magnetic moment, for example due to an external magnetic field in order to align with an external magnetization, always causes a change in angular momentum, too. This is only possible through a mass rotation of the atom. Therefore, an external magnetic field exerts torque and causes a gyration of the atom, hence the term gyromagnetic effect (Waters and Francis, 1958; Stuart, 1972; Abrahams and Keffer, 2019). In the process of realigning the magnetic momentum, the momentum’s axis will start to rotate or process around the external magnetic field direction with a constant frequency (Larmor frequency), which can induce an electromotive force in a pickup-coil. Through the relationship between the Larmor precession frequency, the gyromagnetic ratio, and the magnetic field strength, it is possible to measure the magnetic field by probing the Larmor frequency when the gyromagnetic ratio is known. (Tierney et al., 2019) This is the basis for precession magnetometers (Fig. 8).

Precession magnetometer In order to establish a precession magnetometer, the substance reacting to the measured magnetic field has to possess a net magnetization. This is the case for unbalanced nuclei such as the noble gas isotopes 3He and 129Xe and for protons, for example. Subsequently the devices are denoted nuclear precession or proton precession magnetometers, depending on the probe substance. For a detectable signal, the states of magnetization of the nuclei or protons are required to be polarized as much as possible, with the spin state population in one state exceeding (parallel/antiparallel) the other. In proton precession, the probe contains ample hydrogen atoms and an artificial, strong magnetic field supplied by a coil polarizes the substance. Once the artificial field is turned off, the protons slowly reverse back to align with the ambient field and continue processing around it. When the artificial field is

Fig. 8 Schematic representation of an atom, nucleus or proton with magnetic momentum (A) at zero external field and (B) with the influence of a magnetic field. The magnetic momentum realigns and processes.

Physical Sensors: Magnetic Sensors

107

perpendicular to the external field, the excitation coil is suited to pick up the Larmor frequency signal (Waters and Francis, 1958; Stuart, 1972; Gemmel et al., 2010; Abrahams and Keffer, 2019). Since the sensitivity of nuclear and proton magnetometers depends on the degree of polarization, a high degree of polarization in one state is desirable. In addition to the strong field, for proton magnetometers, it is also possible to increase polarization with the Overhauser effect. The Overhauser effect describes the transfer of an electron’s spin state to a proton. In magnetometers, this can be used through the addition of free and unpaired electrons to the proton sample, thereby achieving a much greater polarization degree with less power. The term Overhauser magnetometer therefore refers to the method of polarization in the device.

Optical atomic magnetometers As described, the change of spin state can be facilitated through the magnetic moment. Besides a large magnetic bias field and the transfer of spins from electrons, this can also be achieved through what is called optical pumping. Here, through energy from polarized laser light, angular momentum is transferred to the probe substance in order to place the substance in a uniform state. Through the linkage of angular to magnetic momentum, optical pumping can completely polarize the probe substance to a single state. This eliminates the need for a very strong excitation field. Optical pumping magnetometers (OPM) are most commonly atomic magnetometers, making use of 4He and alkali metal atomic vapors (K, Cs, Rb); the latter, since the behavior of their single unpaired electron in the outer shell governs the behavior of the vapor atoms (Kominis et al., 2003; Budker and Romalis, 2007; Tierney et al., 2019; Oelsner et al., 2020; Zhang et al., 2020). The advantage compared to SQUID is that OPM systems do not need cooling. The power consumption for the optical polarization, however, is considerably increased (compared to SQUID, not compared to the excitation magnetic field). Optical equipment such as discharge lamps and diode lasers not only enable enhanced magnetic state polarization, it can also probe the resonant precession frequency through the rotation of the plane of polarized light. This has led to the development of all-optical atomic magnetometers with a pump-probe optical setup (Li et al., 2018b; Tierney et al., 2019). The remaining limiting factor for sensitivity is the spin relaxation time of the probe atoms. Contributions of spin-exchange collisions between probe atoms, collisions of probe atoms with the vessel walls, electron spin randomization, the diffusion of atoms and inhomogeneities in the ambient magnetic field influence relaxation. These factors, however, can be controlled and minimized, which leads to the term of SERF magnetometers. The so-called spin exchange relaxation free operation of optical magnetometers refers to the attempt to prolong the spin relaxation time to exceed the Larmor frequency and maintain the polarization of the vapor. One achieves this by using a lower magnetic field and a higher density of the metal (Allred et al., 2002; Kominis et al., 2003; Savukov and Romalis, 2005; Li et al., 2018b). Alkali magnetometers operated in the SERF regime have recently yielded sensitivities lower than the ones observed for SQUID systems (Budker and Romalis, 2007). In conclusion, the field of magnetic sensors is a broad one. For different applications, the last decades brought forth ample sensing concepts. Hall and magnetoresistive sensors, on one hand, are widely used for industrial and automotive applications. GMR and TMR contribution in that particularly field will likely increase, replacing some AMR systems. Flux-Gate magnetometers, gladly combined with search-coils (not covered here), SQUIDs and quantum effect magnetometers, on the other hand, are employed in biomedical applications, geomagnetic applications, and research concerning fundamental material properties and interactions. Not covered here, but very promising for this particular area of applications, nitrogen-vacancy centers in diamonds have been studied intensely in the past few years and will most likely step up to join the ranks of SQUIDs and quantum effect magnetometers and place scanning magnetometry and nuclear magnetic resonance in the nanoscale regime.

References Abrahams, E., Keffer, F., 2019. Gyromagnetic effect. In: Access Science. McGraw-Hill Education . https://doi.org/10.1036/1097-8542.303900. Akouala, C.R., et al., 2019. Planar Hall effect and anisotropic magnetoresistance in semiconducting and conducting oxide thin films. Applied Physics A: Materials Science and Processing 125 (5), 1–12. https://doi.org/10.1007/s00339-019-2592-y. Alam, J., et al., 2020. Cylindrical micro and nanowires: Fabrication, properties and applications. Journal of Magnetism and Magnetic Materials 513, 167074. https://doi.org/ 10.1016/j.jmmm.2020.167074. Ali, M.N., et al., 2014. Large, non-saturating magnetoresistance in WTe2. Nature 514 (7521), 205–208. https://doi.org/10.1038/nature13763. Allred, J.C., et al., 2002. High-sensitivity atomic magnetometer unaffected by spin-exchange relaxation. Physical Review Letters 89 (13), 1308011–1308014. https://doi.org/ 10.1103/physrevlett.89.130801. Andò, B., et al., 2006. Multilayer based technology to build RTD fluxgate. Sensors & Tranducers Magazine 65 (3), 509–514. Andò, B., et al., 2010. RTD fluxgate behavioral model for circuit simulation. Procedia Engineering 5, 1288–1291. https://doi.org/10.1016/j.proeng.2010.09.349. Atulasimha, J., Flatau, A.B., 2011. A review of magnetostrictive iron-gallium alloys. Smart Materials and Structures 20 (4). https://doi.org/10.1088/0964-1726/20/4/043001. Aue, M., 2020. Prozessentwicklung eines Drucksensorsystems auf einem Glassubstrat. Gottfried Wilhelm Leibniz Universität Hannover. Balevicius, S., et al., 2019. Hand-held magnetic field meter based on colossal magnetoresistance-B-scalar sensor. In: IEEE Transactions on Instrumentation and Measurement, IEEE, p. 1. https://doi.org/10.1109/tim.2019.2925411. Baraduc, C., Chshiev, M., Dieny, B., 2013. Spintronic phenomena: Giant magnetoresistance, tunnel magnetoresistance and spin transfer torque. In: Reig, C., Susana Cardoso, D.F., Mukhopadhyay, S.C. (Eds.), Giant Magnetoresistance (GMR) Sensors. Springer Berlin Heidelberg, Heidelberg, pp. 1–30. Barraud, C., et al., 2010. Magnetoresistance in magnetic tunnel junctions grown on flexible organic substrates. Applied Physics Letters 96 (7), 1–4. https://doi.org/10.1063/ 1.3300717. Bason, Y., et al., 2004. Giant planar Hall effect in colossal magnetoresistive La 0.84Sr 0.16MnO3 thin films. Applied Physics Letters 84 (14), 2593–2595. https://doi.org/10.1063/ 1.1695197.

108

Physical Sensors: Magnetic Sensors

Bloom, F.L., et al., 2007. Separating positive and negative magnetoresistance in organic semiconductor devices. Physical Review Letters 99 (25), 1–4. https://doi.org/10.1103/ PhysRevLett.99.257201. Budker, D., Romalis, M., 2007. Optical magentometry. Nature 3, 227–234. Bürgler, D.E., 2010. Spintronics: Magnetism and spin-dependent transport on the nanometer scale. Journal of Siberian Federal University 3 (1), 23–47. Available at: http://www. mathnet.ru/eng/jsfu100. Burkov, A.A., 2017. Giant planar hall effect in topological metals. Physical Review B 96 (4), 1–5. https://doi.org/10.1103/PhysRevB.96.041110. Butvin, P., et al., 2012. Field annealed closed-path fluxgate sensors made of metallic-glass ribbons. Sensors and Actuators A: Physical 184, 72–77. https://doi.org/10.1016/ j.sna.2012.07.001. Calkins, F.T., Flatau, A.B., Dapino, M.J., 2007. Overview of magnetostrictive sensor technology. Journal of Intelligent Material Systems and Structures 18 (10), 1057–1066. https:// doi.org/10.1177/1045389X06072358. Clarke, J., 1989. Principles and applications of SQUIDs. Proceedings of the IEEE 77 (8), 1208–1223. https://doi.org/10.1109/5.34120. Dagotto, E., 2003. In: Cardona, M., et al. (Eds.), Nanoscale Phase Separation and Colossal Magnetoresistance. Springer Berlin Heidelberg, Heidelberg. Damsgaard, C.D., et al., 2008. Exchange-biased planar hall effect sensor optimized for biosensor applications. Journal of Applied Physics 103 (7), 5–8. https://doi.org/10.1063/ 1.2830008. Demirci, E., 2020. Magnetic and magnetotransport properties of memory sensors based on anisotropic magnetoresistance. Journal of Superconductivity and Novel Magnetism. https://doi.org/10.1007/s10948-020-05646-4. Díaz-Michelena, M., 2009. Small magnetic sensors for space applications. Sensors 9 (4), 2271–2288. https://doi.org/10.3390/s90402271. Dixon, R., 2019. A technology roadmap for XMR sensors in automotive applications. In: 15th xMR Symposium, Wetzlar, pp. 1–6. Djamal, M., et al., 2011. Development of fluxgate sensors and its applications. In: International Conference on Instrumentation, Communication, Information Technology and Biomedical Engineering, 2011, pp. 421–426. https://doi.org/10.1109/ICICI-BME.2011.6108590. Ejsing, L., et al., 2004. Planar hall effect sensor for magnetic micro- and nanobead detection. Applied Physics Letters 84 (23), 4729–4731. https://doi.org/10.1063/1.1759380. Ekreem, N.B., et al., 2007. An overview of magnetostriction, its use and methods to measure these properties. Journal of Materials Processing Technology 191 (1–3), 96–101. https://doi.org/10.1016/j.jmatprotec.2007.03.064. Fagaly, R.L., 2006. Superconducting quantum interference device instruments and applications. Review of Scientific Instruments 77 (10). https://doi.org/10.1063/1.2354545. Fina, I., et al., 2014. Anisotropic magnetoresistance in an antiferromagnetic semiconductor. Nature Communications 5, 1–7. https://doi.org/10.1038/ncomms5671. Freitas, P.P., Ferreira, R., Cardoso, S., 2016. Spintronic sensors. Proceedings of the IEEE 104 (10), 1894–1918. https://doi.org/10.1109/JPROC.2016.2578303. Gemmel, C., et al., 2010. Ultra-sensitive magnetometry based on free precession of nuclear spins. European Physical Journal D 57, 303–320. https://doi.org/10.1140/epjd/e201000044-5. Goel, N., et al., 2020. Hall effect instruments, evolution, implications, and future prospects. The Review of Scientific Instruments 91 (7), 071502. https://doi.org/10.1063/ 5.0009647. Gomonay, H., Loktev, V.M., 2002. Magnetostriction and magnetoelastic domains in antiferromagnets. Journal of Physics. Condensed Matter 14 (15), 3959–3971. https://doi.org/ 10.1088/0953-8984/14/15/310. Groenland, J.P.J., et al., 1992. Permalloy thin-film magnetic sensors. Sensors and Actuators, A: Physical 30 (1–2), 89–100. https://doi.org/10.1016/0924-4247(92)80201-D. Guo, L., et al., 2019. Spin transport in organic molecules. Frontiers in Chemistry 7, 1–11. https://doi.org/10.3389/fchem.2019.00428. Henriksen, A.D., et al., 2010. Planar hall effect bridge magnetic field sensors. Applied Physics Letters 97 (1), 2008–2011. https://doi.org/10.1063/1.3460290. Hewett, T.H., Kusmartsev, F.V., 2010. Geometrically enhanced extraordinary magnetoresistance in semiconductor-metal hybrids. Physical Review B: Condensed Matter and Materials Physics 82 (21), 3–6. https://doi.org/10.1103/PhysRevB.82.212404. Hu, B., Wu, Y., 2007. Tuning magnetoresistance between positive and negative values in organicsemiconductors. Nature Materials 6 (12), 985–991. https://doi.org/10.1038/ nmat2034. Janutka, A., Brzuszek, K., 2020. Giant magnetoreactance in magnetic nanowires. Journal of Magnetism and Magnetic Materials 515, 167297. https://doi.org/10.1016/ j.jmmm.2020.167297. Jin, S., et al., 1994. Thousandfold change in resistivity in magnetoresistive La-Ca-Mn-O films. Science 264 (5157), 413–415. https://doi.org/10.1126/science.264.5157.413. Jogschies, L., et al., 2014. Investigations on strain behaviour of polymer substrates during a separation process. Procedia Technology 15, 221–229. https://doi.org/10.1016/ j.protcy.2014.09.075. Johnson, H.G., et al., 2010. Universal properties of linear magnetoresistance in strongly disordered MnAs-GaAs composite semiconductors. Physical Review B: Condensed Matter and Materials Physics 82 (8), 1–4. https://doi.org/10.1103/PhysRevB.82.085202. Joshi, G., et al., 2018. High-field magnetoresistance of organic semiconductors. Physical Review Applied 10 (2), 1. https://doi.org/10.1103/PhysRevApplied.10.024008. Knobel, M., Pirota, K.R., 2002. Giant magnetoimpedance: Concepts and recent progress. Journal of Magnetism and Magnetic Materials 242–245, 33–40. https://doi.org/10.1016/ S0304-8853(01)01180-5. Knudde, S., et al., 2017. Annealing free magnetic tunnel junction sensors. Journal of Physics D: Applied Physics 50 (16). https://doi.org/10.1088/1361-6463/aa622a. Kokado, S., Tsunod, M., 2013. Anisotropic magnetoresistance effect: General expression of AMR ratio and intuitive explanation for sign of AMR ratio. Advanced Materials Research 750–752, 978–982. https://doi.org/10.4028/www.scientific.net/AMR.750-752.978. Kominis, I.K., et al., 2003. A subfemtotesla multichannel atomic magnetometer. Nature 422 (6932), 596–599. https://doi.org/10.1038/nature01484. Krischik, J., Rissing, L., 2015. Funktionsoptimierung von AMR Sensoren auf flexiblem Polyimid. Gottfried Wilhelm Leibniz Universität Hannover. Lee, E.W., 1955. Magnetostriction and magnetomechanical effects. Reports on Progress in Physics 18 (1), 184–229. https://doi.org/10.1088/0034-4885/18/1/305. Leitao, D.C., et al., 2015. Magnetoresistive nanosensors: Controlling magnetism at the nanoscale. Nanotechnology 27 (4), 45501. https://doi.org/10.1088/0957-4484/27/4/ 045501. Li, B., Kavaldzhiev, M.N., Kosel, J., 2015. Flexible magnetoimpedance sensor. Journal of Magnetism and Magnetic Materials 378, 499–505. https://doi.org/10.1016/ j.jmmm.2014.11.067. Li, H., et al., 2018a. Giant anisotropic magnetoresistance and planar hall effect in the Dirac semimetal Cd3As2. Physical Review B 97 (20), 1–6. https://doi.org/10.1103/ PhysRevB.97.201110. Li, J., et al., 2018b. SERF atomic magnetometer-recent advances and applications: A review. IEEE Sensors Journal 18 (20), 8198–8207. https://doi.org/10.1109/ JSEN.2018.2863707. Liu, J.H., Jiang, C.B., Xu, H.B., 2012. Giant magnetostrictive materials. SCIENCE CHINA Technological Sciences 55 (5), 1319–1326. https://doi.org/10.1007/s11431-012-4810-0. Liu, L., et al., 2016. Magnetostrictive GMR spin valves with composite FeGa/FeCo free layers. AIP Advances 6 (3). https://doi.org/10.1063/1.4943770. Lu, M.W., et al., 2016. Controllable giant magnetoresistance effect by the d-doping in a magnetically confined semiconductor heterostructure. Applied Surface Science 360, 989– 993. https://doi.org/10.1016/j.apsusc.2015.11.101. Lühr, H., et al., 1985. The IRM fluxgate magnetometer. IEEE Transactions on Geoscience and Remote Sensing GE-23 (3), 259–261. https://doi.org/10.1109/TGRS.1985.289524. Mell, H., Stuke, J., 1970. Negative magnetoresistance of amorphous semiconductors. Journal of Non-Crystalline Solids 4, 304–310. https://doi.org/10.1016/0022-3093(70) 90055-4. Nichterwitz, M., et al., 2020. Control of positive and negative magnetoresistance in iron oxide-iron nanocomposite thin films for tunable magnetoelectric nanodevices. ACS Applied Electronic Materials 2 (8), 2543–2549. https://doi.org/10.1021/acsaelm.0c00448.

Physical Sensors: Magnetic Sensors

109

Nikitin, A., Stocks, N.G., Bulsara, A.R., 2013. Bistable sensors based on broken symmetry phenomena: The residence time difference vs. the second harmonic method. European Physical Journal: Special Topics 222 (10), 2583–2593. https://doi.org/10.1140/epjst/e2013-02039-2. Oelsner, G., et al., 2020. Integrated optically pumped magnetometer for measurements within Earth’s magnetic field. arXiv 1–11. Palsule S, Sortee S and Jayamon J (2019) Magnetic Sensor Market Size, Share & Trends Analysis Report By Technology, By Application, By Region, And Segment Forecasts, 2019–2025. Pang, N., Cheng, D., Wang, Y., 2017. A high stability time difference readout technique of RTD-fluxgate sensors. Sensors 17 (10), 1–15. https://doi.org/10.3390/s17102325. Parkin, S.S.P., et al., 2004. Giant tunnelling magnetoresistance at room temperature with MgO (100) tunnel barriers. Nature Materials 3 (12), 862–867. https://doi.org/10.1038/ nmat1256. Pigeau, B., 2012. Magnetic Vortex Dynamics Nanostructures. Université Paris Sud, Paris, p. IX. Pippard, A.B., 1989. Magnetoresistance in Metals. Cambridge University Press, Cambridge. Popovic, R.S., Flanagan, J.A., Besse, P.A., 1996. The future of magnetic sensors. Sensors and Actuators A: Physical 56 (1–2), 39–55. https://doi.org/10.1016/0924-4247(96) 01285-X. Prediger, M., Rechel, M., Wurz, M.C., 2019a. AMR-based mechanical pressure sensors: A proof of concept. In: 15th xMR Symposium. Prediger, M.S., et al., 2019b. AMR Sensoren auf flexiblem Substrat zur Messung von Verformun- gen eines hartmagnetischen Partikel-Elastomerverbunds. In: MikroSystemTechnik Kongress. VDE Verlag GmbH, Berlin, pp. 111–114. Primdahl, F., 1979. The fluxgate magnetometer. Journal of Physics E: Scientific Instruments 12, 241. Available at: http://iopscience.iop.org/article/10.1088/0022-3735/12/4/ 001/pdf. Prinz, G.A., 1998. Magnetoelectronics. Science 282, 1660–1663. Ramirez, A.P., 1997. Colossal magnetoresistance. Journal of Physics. Condensed Matter 9, 8171–8199. Reig, C., Cubells-Beltran, M.D., 2013. GMR based sensors for IC current monitoring. In: Reig, C., De Freitas, S.C., Mukhopadhyay, S.C. (Eds.), Giant Magnetoresistance (GMR) Sensors. Springer Berlin Heidelberg, Heidelberg, pp. 103–132. Reig, C., De Freitas, S.C., Mukhopadhyay, S.C., 2013. Giant Magnetoresistance (GMR) Sensors. Springer Berlin Heidelberg, Heidelberg. Ren, L., Yu, K., Tan, Y., 2019. Applications and advances of magnetoelastic sensors in biomedical engineering: A review. Materials 12 (7). https://doi.org/10.3390/ma12071135. Ripka, P., 1992. Review of fluxgate sensors. Sensors and Actuators, A: Physical 33 (3), 129–141. https://doi.org/10.1016/0924-4247(92)80159-Z. Ripka, P., 1994. Magnetic sensors for industrial and field applications. Sensors and Actuators, A: Physical 42 (1–3), 394–397. https://doi.org/10.1016/0924-4247(94)80019-7. Ripka, P., Arafat, M.M., 2019. Magnetic sensors: Principles and applications. Reference Module in Materials Science and Materials Engineering 1, 1–15. https://doi.org/10.1016/ b978-0-12-803581-8.11680-7. Rittinger, J., et al., 2015. Impact of different polyimide-based substrates on the soft magnetic properties of NiFe thin films. Smart Sensors, Actuators, and MEMS VII; and Cyber Physical Systems 9517, 95171R. https://doi.org/10.1117/12.2178765. Savukov, I.M., Romalis, M.V., 2005. Effects of spin-exchange collisions in a high-density alkali-metal vapor in low magnetic fields. Physical Review A: Atomic, Molecular, and Optical Physics 71 (2), 2–9. https://doi.org/10.1103/PhysRevA.71.023405. Schuhl, A., Van Dau, F.N., Childress, J.R., 1995. Low-field magnetic sensors based on the planar hall effect. Applied Physics Letters 2751, 2751. https://doi.org/10.1063/ 1.113697. Schüthe, T., Albounyan, A., Riemschneider, K.R., 2019. Two-dimensional characterization and simplified simulation procedure for tunnel magnetoresistive angle sensors. In: Conference Proceedings SAS 2019–2019 IEEE Sensors Applications Symposium. https://doi.org/10.1109/SAS.2019.8706125. Shumilin, A.V., 2020. Microscopic theory of organic magnetoresistance based on kinetic equations for quantum spin correlations. Physical Review B: American Physical Society 101 (13), 1–14. https://doi.org/10.1103/PhysRevB.101.134201. Solin, S.A., et al., 2000. Enhanced room-temperature geometric magnetoresistance in inhomogeneous narrow-gap semiconductors. Science 289 (5484), 1530–1532. https:// doi.org/10.1126/science.289.5484.1530. Storm, J.H., et al., 2017. An ultra-sensitive and wideband magnetometer based on a superconducting quantum interference device. Applied Physics Letters 110 (7). https://doi.org/ 10.1063/1.4976823. Stuart, W.F., 1972. Earth’s field magnetometry. Reports on Progress in Physics 35 (2), 803–881. https://doi.org/10.1088/0034-4885/35/2/306. Sun, J., Kosel, J., 2013. Extraordinary magnetoresistance in semiconductor/metal hybrids: A review. Materials 6 (2), 500–516. https://doi.org/10.3390/ma6020500. Tang, H.X., et al., 2003. Giant planar hall effect in epitaxial (Ga, Mn)As devices. Physical Review Letters 90 (10), 4. https://doi.org/10.1103/PhysRevLett.90.107201. Tanty, N., et al., 2019. Tuning magnetoresistance and electrical resistivity by enhancing localization length in polyaniline and carbon nanotube composites. Bulletin of Materials Science 42 (5), 1–7. https://doi.org/10.1007/s12034-019-1890-5. Tian, Q., Xie, S., 2019. Spin injection and transport in organic materials. Micromachines 10 (9). https://doi.org/10.3390/mi10090596. Tierney, T.M., et al., 2019. Optically pumped magnetometers: From quantum origins to multi-channel magnetoencephalography. NeuroImage 199, 598–608. https://doi.org/ 10.1016/j.neuroimage.2019.05.063. Trabaldo, E., et al., 2020. SQUID magnetometer based on Grooved Dayem nanobridges and a flux transformer. IEEE Transactions on Applied Superconductivity 30 (7), 1–4. https:// doi.org/10.1109/TASC.2020.3004896. Trigona, C., et al., 2018. Measurements of Iron compound content in the brain using a flexible Core fluxgate magnetometer at room temperature. IEEE Transactions on Instrumentation and Measurement 67 (4), 971–980. https://doi.org/10.1109/TIM.2018.2791218. Trigona, C., et al., 2020. RTD-fluxgate magnetometers for detecting iron accumulation in the brain. IEEE Instrumentation and Measurement Magazine 23 (1), 7–13. https://doi.org/ 10.1109/MIM.2020.8979517. Tu, C., et al., 2019. Mechanical-resonance-enhanced thin-film magnetoelectric heterostructures for magnetometers, mechanical antennas, tunable RF inductors, and filters. Materials 12 (14). https://doi.org/10.3390/ma12142259. Tumanski, S., 2013. Modern magnetic field sensorsdA review. Przeglad Elektrotechniczny 10, 1–12. Available at: http://www.tumanski.pl/01_PE_10_13_1-12_tumanski.pdf. Vaidhyanathan, L.S., Baisnab, D.K., Janawadkar, M.P., 2011. Development of high Tc superconducting quantum interference devices. AIP Conference Proceedings 1349 (Part A), 463–464. https://doi.org/10.1063/1.3605934. Vidal, E.G., et al., 2017. Electronic energy meter based on a tunnel magnetoresistive effect (TMR) current sensor. Materials 10 (10), 1–11. https://doi.org/10.3390/ma10101134. Vincent, J.D.S., et al., 2020. Design and development of magnetostrictive actuators and sensors for structural health monitoring. Sensors (Switzerland) 20 (3). https://doi.org/ 10.3390/s20030711. Wadhwani P, Global MI and Yadav S (2019) Magnetic Sensor Market Size By Technology, By Application, By End-Use, Industry Analysis Report, Regional Outlook, Application Development Potential, Price Trends, Competitive Market Share & Forecast, 2019–2025. Wadhwani, P, Global M. I and Yadav, S (2020) Magnetic Sensor Market Size By Technology, By Application, By End-Use Industry Analysis Report, Regional Outlook, Application Development Potential, Price Trend, Competitive Market Share & Forecast, 2020–2026. Wagemans, W., Koopmans, B., 2011. Spin transport and magnetoresistance in organic semiconductors. Physica Status Solidi B: Basic Research 248 (5), 1029–1041. https:// doi.org/10.1002/pssb.201046383. Waters, G.S., Francis, P.D., 1958. A nuclear magnetometer. Journal of Scientific Instruments 38, 88–93. Weitensfelder, H., Brueckl, H., Satz, A., Pruegl, K., et al., 2018a. Comparison of sensitivity and low-frequency noise contributions in Giant-Magnetoresistive and tunnelingmagnetoresistive spin-valve sensors with a vortex-state free layer. Physical Review Applied 10 (5), 1. https://doi.org/10.1103/PhysRevApplied.10.054056.

110

Physical Sensors: Magnetic Sensors

Weitensfelder, H., Brueckl, H., Satz, A., Suess, D., 2018b. Noise characterization of vortex-state GMR sensors with different free layer thicknesses. Proceedings 2 (13), 1013. https://doi.org/10.3390/proceedings2131013. Wittek, C.-G.R., 2019. Prozessentwicklung von AMR Sensoren auf Flexiblem Substrat mit Durchkontaktierung. Gottfried Wilhelm Leibniz Universität Hannover. Wurft T (2018) Investigation of the Magnetic Vortex State for Spin-Valve Sensors. Available at: https://pub.uni-bielefeld.de/record/2931263. Wurft, T., et al., 2017. The influence of edge Inhomogeneities on vortex hysteresis curves in magnetic tunnel junctions. IEEE Transactions on Magnetics 53 (11). https://doi.org/ 10.1109/TMAG.2017.2715072. Yang, Z., et al., 2019. Patterned soft ferromagnetic meander-shaped elements for sensor application. Sensors 19 (2468), 1–12. Zhang, X., et al., 2018. Fabrication and measurement of Nb-based SQUID magnetometer. Physica C: Superconductivity and Its Applications 548, 1–4. https://doi.org/10.1016/ j.physc.2018.01.016. Zhang, R., et al., 2020. Recording brain activities in unshielded Earth’s field with optically pumped atomic magnetometers. Science Advances 6 (24), 1–9. https://doi.org/10.1126/ sciadv.aba8792. Zheng, C., et al., 2019. Magnetoresistive sensor development roadmap (non-recording applications). IEEE Transactions on Magnetics 55 (4). https://doi.org/10.1109/ TMAG.2019.2896036. Zhukov, A., Ipatov, M., Zhukova, V., 2015. Advances in Giant magnetoimpedance of materials. In: Handbook of Magnetic Materials. Elsevier Ltd. https://doi.org/10.1016/ bs.hmm.2015.09.001. Zhukov, A., et al., 2020. Giant magnetoimpedance in rapidly quenched materials. Journal of Alloys and Compounds 814. https://doi.org/10.1016/j.jallcom.2019.152225. Zhukova, V., et al., 2018. Tailoring of magnetoimpedance effect and magnetic softness of Fe-rich glass-coated microwires by stress-annealing. Scientific Reports 8 (1), 1–14. https://doi.org/10.1038/s41598-018-21356-3.

Relevant Websites www.akm.com/global/endTechnologydTechnical TutorialdBasic Knowledge of Magnetic Sensor. www.electronics-tutorials.wsdElectromagnetismdBasic Knowledge and Hall Effect Sensor.

Physical Sensors: Motion Sensors for Physical Activity Assessment Jeffer Eidi Sasakia and Robert W. Motlb, a Department of Sport Sciences, Federal University Triângulo Mineiro, Uberaba, Brazil; and b Department of Kinesiology and Nutrition, College of Applied Health Sciences, University of Illinois Chicago, Chicago, IL, United States © 2023 Elsevier Ltd. All rights reserved.

Introduction Brief history of motion sensors Pedometers Technical aspects Validity of pedometers Stepping rate and moderate-to-vigorous intensity physical activity Motivational device Steps per day recommendations Accelerometers Technical aspects Estimating physical activity from acceleration signals Linear regression equations Machine learning algorithms Parameters of device use Multisensor and posture-based activity monitors Consumer-grade accelerometers/fitness trackers What is next? Summary References

111 112 112 112 112 114 114 114 115 115 115 115 116 116 118 119 119 120 120

Glossary Accelerometer Sensor that measures acceleration. Light intensity physical activity (LPA) Activities with low-energy expenditure, with MET values ranging between 1.5 and 2.99 METs. Examples: slow walking, cooking, washing dishes. Metabolic equivalent of task (MET) Is a unit for expressing the rate of energy expenditure of activities. One MET is defined as the energy expenditure while resting and is approximately 3.5 mL of O.2kg 1.min 1. Moderate-to-vigorous intensity physical activity (MVPA) Activities with an intensity  3 METs and < 6 METs. A common example of a MVPA is walking. Pedometer Motion sensor that counts steps, distance traveled, and kcals burned. Physical activity Any bodily movement produced by the skeletal muscles that result in energy expenditure. Validity The degree to which an instrument measures what is intended to measure. Vigorous intensity physical activity (VPA) Activities with high-energy expenditure, with MET values  6 METs. Examples: climbing stairs, running, sports.

Abstract Motion sensors have been increasingly used in physical activity research. These devices have the potential to improve the accuracy of physical activity estimates in free-living studies and, consequently, help researchers in grasping a better understanding of the association of physical activity with health outcomes. In this article, we present an overview of the applications of motion sensors for physical activity assessment, as well as some background on the scientific knowledge generated over the last two decades.

Introduction Physical activity has been defined as any bodily movement produced by contraction of the skeletal muscles that results in energy expenditure (Caspersen et al., 1985). This behavior has been of major interest in different scientific disciplines because of its

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00056-X

111

112

Physical Sensors: Motion Sensors for Physical Activity Assessment

relationship with health in humans (Warburton et al., 2006). The scientific study of physical activity and health has been advanced as researchers have developed tools that assess physical activity in the field, although much of the research has relied on surveys and questionnaires. The advancements in technology resulting in motion sensors seemingly provided more accurate device-based assessment of free-living physical activity. Motion sensors have seemingly improved the accuracy of physical activity measures by avoiding recall/memory biases and over-reporting and, thus, present several important applications for research, including: (a) better identification of physical activity levels in free-living settings, (b) elucidation of dose-response relationships between physical activity and health outcomes, and (c) enhanced monitoring of physical activity behavior change in interventions. Despite these potential benefits and applications, researchers should be aware of several key issues before making decisions involving the use of these devices. Among these issues, researchers should understand the features and technology of the different motion sensors, the appropriate form of utilization in the field, the methods available for data processing, the validity evidence, new possibilities of use, as well as the limitations of the existing motion sensors. This article provides an overview of the use and applications of motion sensors for physical activity assessment. These devices are usually of two main types, pedometers or accelerometer-based activity monitors, and this article will primarily focus on these two types of motion sensors. For clarity, we will refer to accelerometer-based activity monitors as only accelerometers throughout the manuscript, although these devices also provide other outputs besides acceleration signals.

Brief history of motion sensors Objective monitoring of physical activity likely evolved from the initial invention of a step counter by Leonardo Da Vinci. The device was worn on the waist and a step was registered by the rotation of gears through the movement of a long lever arm attached to the thigh (Gibbs-Smith and Rees, 1978). Thomas Jefferson in the late 1780s requested that a watchmaker from Paris build a small-size step counter. The device could fit in a vest pocket, and had a lever arm connected to a string with one of the ends strapped right below the knee. Leg movements would cause the lever arm to move gears and produce the registration of steps. The original intention of Thomas Jefferson was to measure the steps taken (i.e., distance traveled) while walking to Paris Landmarks (Dumbauld, 1946). Almost two centuries later, the Japanese Company Yamasa (Yamax) developed the famous Manpo-Kei (in English “10,000 steps meter”). The recommendation of 10,000 steps/day for health benefits dates back to 1965 when the Manpo-Kei was released in the market (Bassett et al., 2017). Throughout the years, Yamax has produced several different models of pedometers, with the majority of them considered among the most accurate devices. Researchers have frequently accepted the device developed by Montoye et al. (1983) as the first accelerometer, and this later resulted in the development of the well-known Caltrac activity monitor; the Caltrac was the predominant accelerometer among researchers for nearly a decade. The Computer Science and Applications (CSA) activity monitor developed in 1995 gained the attention from researchers and benefited from studies examining its validity. The popularization of the CSA activity monitor became more evident after calibration studies developed methods for predicting energy expenditure-related variables from the CSA output. The CSA brand later became ActiGraph and this brand of activity monitors has become the most widely used by the research community. Along with the ActiGraph activity monitors, several others have been developed and used in research, such as the RT3 and RT6, Actical, and GeneActiv.

Pedometers Technical aspects Pedometers are small electronic devices, usually worn on the waist, and are used for measuring the number of steps taken (i.e., binary outcome) during a given period of time (e.g., minute, hour, day, week) (Tudor-Locke et al., 2012). Many devices are currently available in the market, and these present with different sensing technologies and features. The simplest pedometers use a springlevered system as the sensing mechanism, by which up-and-down movements result in a circuit closing and, thus, the registration of steps (Bassett et al., 2017) (Fig. 1). On the other hand, newer pedometers use micromachined electromechanical systems (MEMS) that detect acceleration based on the magnitude of voltage output generated by the sensing mechanism (e.g., differential capacitance) (Fig. 1). The acceleration is then transformed into steps by means of proprietary algorithms (Bassett et al., 2017). Some pedometers may present memory function for recording step data over multiple days and other measures such as kcals (i.e., energy expenditure), distance traveled, and aerobic steps. Some pedometers may even include an interface software for downloading and analyzing data. Of note, current accelerometers (e.g., ActiGraph) provide measures of steps and may be used in some studies as pedometers. Smartphones nowadays support pedometer apps with great interfaces for monitoring steps as well as kcals (i.e., energy expenditure) and distance traveled, and the data can even be shared with one’s circle of friends or social media account.

Validity of pedometers Studies have indicated that accuracy of step counting is influenced by the sensing mechanism of the device, as well as locomotion speed, activity type, movement pattern, and weight status of users. These studies have tested the validity of waist-worn pedometers against criterion measures such as manual counting and the StepWatch (highly accurate pedometer worn on the ankle).

Physical Sensors: Motion Sensors for Physical Activity Assessment

Fig. 1

113

Spring-levered and differential capacitance mechanisms for recording steps.

One study by Crouter et al. (2003) revealed that different pedometers significantly underestimated step taken during slow walking speeds on a treadmill. This was likely due to the low magnitude of vertical displacement (up-and-down movement) of the hip in these activities and the internal mechanism of the devices (most were spring-levered pedometers). Measurement errors at the slowest speed of 54 m/min were > 10% for eight of the 10 pedometers. However, six of the ten pedometers were within  1% of actual steps with faster speeds (80, 94, and 107 m/min) (Crouter et al., 2003). The accuracy of pedometers has been tested during self-paced overground walking (Schneider et al., 2003). The measurement error with overground walking was dependent on the pedometer model, with only three of the 10 pedometers measuring steps within  3% of actual steps taken (Schneider et al., 2003). Of note, two of these devices were piezoelectric pedometers (Kenz Lifecorder and New Lifestyles NL-2000), and both are known for their better sensitivity during slow walking speeds. One subsequent study verified that a piezoelectric pedometer (New Lifestyles NL-2000) was more accurate than a spring-levered pedometer (Yamax Digiwalker SW-200) in registering steps during walking speeds of 54, 67, 80, 94 and 107 m/min (Crouter et al., 2005). The study further indicated that the piezoelectric pedometer was more accurate than the spring-levered pedometer for recording steps in overweight individuals during the five-abovementioned speeds. The measurement error was dependent on the tilting angle, with the piezoelectric pedometer undercounting steps by 0.1  11.8%, 0.5  4.6% and 2.6  10.6%, when tilting angles were  10o, 10.1o to 15o, and > 15o, respectively. In contrast, the spring-levered pedometer undercounted steps by 5.6  25.4%, 14.4  28.9% and 41.3  44.1%, when tilting angles were  10o, 10.1o to 15o, and > 15o, respectively (Crouter et al., 2005). Importantly, the tilt angles reflected the influence of overweight status such as greater tilt angles occurred with greater obesity based on presence of abdominal adiposity. In more recent years, researchers have asked the question of “What is a step?” The definition of a step may be debatable, but it is recommended that, for physical activity assessment purposes, all types of steps be captured by the pedometer (Bassett et al., 2017). This includes forward steps, backward steps, side-to-side steps, diagonal steps, puttering steps, walking and running steps. Such understanding is important, considering that for accurate physical activity assessment, pedometers should capture steps during locomotion as well as during several lifestyle and sport activities. In this regard, a study by Hickey et al. (2015) demonstrated that the Yamax SW-200 and the Omron HJ720-ITC pedometers were highly inaccurate in capturing steps from different lifestyle activities (% errors: 41.6% and 79.0%, respectively). Yet with locomotion activities (overground self-paced walking and four treadmill speeds), the overall measurement errors were 8.2% and 11.1% for the Yamax and the Omron pedometers, respectively (Hickey et al., 2015). Researchers have examined the accuracy of pedometers in special populations, such as individuals with neurological conditions, who usually present gait abnormalities and/or walking limitations. One early study by Motl et al. (2005) revealed that springlevered pedometers were accurate (measurement error  4.4%) when recording steps at comfortable and fast walking speeds (i.e., 67, 80, and 94 m/min) in individuals with multiple sclerosis. However, measurement error of such devices was significantly high for slow walking speeds (i.e., 41 and 54 m/min), amounting to up to 31.6% error (Motl et al., 2005). Another study compared step counts recorded by a spring-levered pedometer (i.e., Yamax SW-200) in healthy individuals versus individuals with neurological conditions (Elsworth et al., 2009). The results demonstrated that, during self-selected speeds, the measurement error for steps taken was only about 3% in healthy controls, but varied between 24% and 30% in those with neurological conditions (Elsworth et al., 2009). According to the authors, this difference in measurement error was not related to walking speed. Therefore, it is

114

Physical Sensors: Motion Sensors for Physical Activity Assessment

possible that movement patterns (e.g., shuffling, limping) that minimize displacement of the center of mass in neurological conditions may result in differential registration of steps by spring-levered pedometers. Considering the evidence, researchers and health professionals should be aware that spring-levered pedometers are not adequate for registering steps during slow walking speeds and in persons with altered gait, such as those with neurological conditions. For these cases, the use of piezoelectric or piezoresistive pedometers, which present higher sensitivity for slow walking and low magnitude vertical displacement, are advisable.

Stepping rate and moderate-to-vigorous intensity physical activity The number of steps taken in a given interval, defined as step-rate or cadence, is usually reflective of the effort or intensity demanded by the locomotion activity. Researchers have sought to determine step-rate thresholds for classifying moderate-tovigorous intensity physical activity (MVPA), which are those activities with a metabolic demand of 3 METs or greater (one MET being equal to the energy requirement while resting,  3.5 ml of O2/kg/min or 1 kcal/kg/h). Several of the currently available pedometers are able to register minute-by-minute steps, and therefore providing a feasible way for assessing physical activity intensity in the field. Studies have indicated that a step-rate threshold of approximately 100 steps/min (3000 steps in 30 min) is representative of MVPA, and the step-rate threshold for vigorous intensity physical activity (VPA) is about 130 steps/min in healthy adults (Tudor-Locke et al., 2019). Among older adults, step-rates for MVPA and VPA are slightly higher than in adults, probably because of lower stride lengths. One study by O’Brien et al. (2018) revealed that step-rate thresholds for absolute and relative MVPA ( 3 METs, and 40% of METMax, respectively) in older adults were 108 and 117 steps/min, respectively. For absolute and relative VPA ( 6 METs, and 60% of METMax, respectively), the step-rate thresholds were 135 and 132 steps/min, respectively (O’Brien et al., 2018). Step-rate thresholds for MVPA have further been developed for people with special conditions. Agiovlasitis et al. (2016) observed that step-rate thresholds for MVPA in people with multiple sclerosis differed according to disability status (mild, moderate, or severe) and height. Among those with mild, moderate, and severe disability, step-rate thresholds for MVPA ranged between 92 and 107, 81–96, and 70–85 steps/min, respectively, according to maximum (186 cm) and minimum values (154 cm) of height (Agiovlasitis et al., 2016). Our group recently developed a step-rate threshold for MVPA in people with Parkinson’s disease (Jeng et al., 2020a). That study included 30 individuals with Parkinson’s disease and 30 healthy controls who performed a 6-min bout of overground walking at normal comfortable speed, and three, 6-min bouts of treadmill walking at slower ( 13.4 m/min), comfortable, and faster than comfortable speeds (þ13.4 m/min). The number of steps was manually counted and oxygen consumption was measured by a portable indirect calorimetry system. The results revealed that the MVPA step-rate threshold for individuals with Parkinson’s disease was  80 steps/min, while for healthy matched controls, the step-rate was  93 steps/min (Jeng et al., 2020a).

Motivational device Pedometer readings usually serve as a motivational factor for physical activity behavior change. This may be explained by the Social Cognitive Theory (SCT) principles of self-monitoring and goal setting as well as performance accomplishment that pedometers can facilitate. Indeed, pedometers provide on-going feedback regarding physical activity behavior for self-monitoring one’s status and progress towards goals, and the satisfaction of one’s goals provides mastery experience for boost self-efficacy as a major SCT influence of behavior (Bandura, 2004). There have been several studies demonstrating significant effects of pedometer utilization on positive changes in physical activity (Bravata et al., 2007), and pedometers may help individuals to attain step recommendations for health benefits from a public health perspective.

Steps per day recommendations The concept of a recommended number of steps/day for health has been around since the first commercially available pedometer, the Manpo-Kei. The very name Manpo-Kei means 10,000 steps/day. Based on this initial number and descriptive data from the literature, as well as perceptions of activity levels in different groups, researchers have proposed a classification system for steps/day in healthy adults, where (a) sedentary lifestyle: < 5000 steps/day; (b) physically inactive: 5000–7499 steps/day; (c) moderately active: 7500–9999 steps/day; (d) physically active: 10,000–12,499 steps/day; and (e) very active:  12,500 (Tudor-Locke and Bassett, 2004). However, in a series of three papers, Tudor-Locke et al. (2011a,b,c) observed that the cut-point of 10,000 steps/day for classifying individuals as “physically active” was not aligned with the necessary volume for achieving the physical activity recommendations. Therefore, after examining the literature, the authors proposed new ranges of steps/day for classifying individuals as physically active. The cut-points for each age group were based on the number of steps/day necessary for reaching the recommendations of physical activity. These cut-points are displayed in Table 1. It is important to note that higher steps/day are associated with more favorable cardiometabolic risk profiles.

Physical Sensors: Motion Sensors for Physical Activity Assessment Table 1

115

Minimum number of steps per day for achieving the recommendations of physical activity.

Group

Steps per day a

Children Adolescents Adults Older adults Individuals living with disability or chronic illness

13,000–15,000 10,000–11,700 7000–10,000 7000–8000 6500–8500

a

Tudor-Locke et al. (2011a), Tudor-Locke et al. (2011b), Tudor-Locke et al. (2011c) and TudorLocke and Bassett (2004).

Accelerometers Technical aspects Over the past decade, most accelerometer-based physical activity monitors have been using different micromachined electromechanical systems that detect changes in acceleration of the human body (i.e., bodily movement). These technologies include piezoelectric, capacitive, and piezoresistive sensors (Chen et al., 2012; Chen and Bassett Jr, 2005; Sasaki et al., 2016a). The output of accelerometers is usually provided as raw acceleration (e.g., multiples of g force) or counts; the counts are proprietary measures directly related to the magnitude and frequency of acceleration (Sasaki et al., 2016a). Older accelerometers would only assess these outputs for the vertical axis; however, current accelerometers capture acceleration from the three orthogonal axes (i.e., vertical, antero-posterior, and medio-lateral axes) and some will even capture signals from nine axes (triaxial gyroscope þ triaxial accelerometer þ triaxial magnetic field). Regarding detection range, the latest accelerometers can sense acceleration within  8Gs (i.e., 8 times the gravitational force), which is more than sufficient to capture the vast majority of human movements. Another aspect of importance is the sampling rate capabilities of the current accelerometers. Some of them can sample signals up to 100 Hz, certainly exceeding the needs of any researcher working with physical activity assessment (John et al., 2012). Most accelerometers are of small size and present long lasting battery life and memory (e.g., 4 GB), and some models record data for over 20 consecutive days at a sampling rate of 100 Hz. Collectively, current accelerometers have outstanding capabilities for assessing free-living physical activity during extended periods of time. These capabilities have even outpaced the speed at which researchers develop new methods and techniques for predicting physical activity from accelerometer data.

Estimating physical activity from acceleration signals The sensor technologies used in the accelerometer-based physical activity monitors are not different than those used in aviation, construction, automotive, and industrial fields. The difference lies on the application of the signals obtained with the sensors; in other words, the translation of the signals into physical activity estimates. This process is called value calibration. Value calibration is the process by which a relationship is established between two measures (Bassett Jr et al., 2012). One is known as the predicting variable and the other as the criterion measure. The latter is the variable to be predicted, thus, it is usually obtained with a highly accurate and valid instrument. The predicting variable, on the other hand, is usually an easier or more convenient measure to obtain. Therefore, the aim of the value calibration process is to generate a method with greater feasibility and applicability in the field, at the expense of the accuracy of the measure. Within this context, the predicting variable is activity counts from the accelerometer and the criterion variable is energy expenditure measured from a device often involving indirect calorimetry (i.e., metabolic system), and the data are subjected into linear regression equations.

Linear regression equations Within the context of linear regression equations for physical activity estimation, researchers usually engage participants with a protocol involving different activities while using a metabolic system (criterion measure) and an accelerometer placed on some part of the body, typically the waist or wrist. The acceleration data, usually in activity counts per minute, are then regressed on metabolic data, typically MET values (mL/kg/min of O2 consumed), and this yields a prediction equation whereby energy expenditure is predicted from counts (Sasaki et al., 2011, 2016a); this calibrates the output from accelerometers with energy expenditure as a gold standard of movement intensity. One of the pioneer studies using this approach was that by Freedson et al. (1998), where 50 adults performed three bouts of walking/running on the treadmill across varying speeds (i.e., 4.8, 6.4, and 9.7 km h 1) while wearing an accelerometer on the waist and a facemask attached to a metabolic system. A regression equation for predicting METs from counts was derived and the researchers went further to develop cut-points by solving the equation with the MET value thresholds for light intensity physical activity (< 3.0 METs), moderate intensity physical activity (3.00–5.99 METs), and vigorous intensity physical activity ( 6.00 METs) (Freedson et al., 1998). The cut-point method provided researchers an off-the-shelf resource for classifying physical

116

Physical Sensors: Motion Sensors for Physical Activity Assessment

activity intensity on a minute-by-minute basis. This method has been used extensively in the past two decades and has allowed significant progress in the physical activity assessment field. Several other calibration studies have followed and developed prediction equations and cut-points for classification of physical activity intensity in children and adolescents, adults, older adults and special populations (Copeland and Esliger, 2009; Evenson et al., 2008; Jeng et al., 2020b; Miller et al., 2010; Motl et al., 2009; Sandroff et al., 2014; Sasaki et al., 2011; Swartz et al., 2000; Treuth et al., 2004). After the pioneer study by Freedson et al. (1998), studies included different activities in the calibration protocol, including locomotion activities, sports, and a variety of lifestyle activities (Hendelman et al., 2000; Swartz et al., 2000). Table 2 depicts a few of the vast number of prediction methods available in the literature, along with a summary of the protocol and validity evidence. In general, protocols involving only locomotion activities usually presented higher correlations between accelerometer data and metabolic data. However, the premise of researchers for including sports and lifestyle activities, especially in calibration studies in children, adolescents and adults, was the intermittent nature and upper-limb movement predominance of some of those activities, which are commonly performed in real world situations. Nevertheless, the accuracy of the prediction models is naturally greater for the specific activities included in the calibration protocol (Crouter et al., 2006). For children and adolescents, researchers have recognized the importance of classifying physical activity intensity at intervals shorter than 1 min. For instance, several cut-points were developed for classification of 15 s and 30 s epochs (Romanzini et al., 2012). This approach has been taken because children and adolescents perform short bursts of activities interspersed with rest periods more often than adults and older adults. One segment of the population that has received less attention regarding the development of physical activity prediction methods from accelerometers is that comprised of special groups. Some studies have developed activity count cut-points specific to conditions such as multiple sclerosis (Motl et al., 2009; Sandroff et al., 2012), Parkinson’s disease (Jeng et al., 2020b), cerebral palsy (Clanchy et al., 2011), spinal cord injury (Holmlund et al., 2020), among others. These methods are important given that several conditions manifest in functional limitations that often influence gait (e.g., step length, stride rate, or asymmetric step patterns) that in turn influences the relationship between counts and energy expenditure (Jeng et al., 2020b; Motl et al., 2009; Sandroff et al., 2014).

Machine learning algorithms Over the past decade, much effort has been directed to developing machine learning algorithms for pattern recognition of accelerometry signals and its classification of activity type and perhaps improving the estimates of physical activity intensity. One of the first studies to take this approach in the physical assessment field was that by Staudenmayer et al. (2009). This study had participants perform pre-defined routines of activities, including sedentary, locomotion, lifestyle, and sport activities. While performing the activities, participants wore an ActiGraph 7164 activity monitor and a portable metabolic system. Artificial neural networks were developed and tested for predicting activity type and METs. The models were highly accurate for both types of predictions (activity type: 88.8% of the time; METs root mean squared error: 1.22 METs [95%CI: 1.14–1.30]) (Staudenmayer et al., 2009). Other researchers have also focused on developing and testing prediction models based on machine learning algorithms (DeVries et al., 2011; Freedson et al., 2011; Mannini et al., 2013; Trost et al., 2012). These studies have demonstrated high accuracy in classifying activity type and predicting activity intensity in laboratory settings. The recent use of raw acceleration data instead of counts brought the possibility of improving even further the classification of activity type, as the resolution of raw acceleration data is substantially higher than counts. Overall, studies utilizing machine learning algorithms for classifying physical activity type from raw acceleration data have performed well in laboratory settings, with some achieving almost perfect accuracy (Zhang et al., 2012a,b). However, researchers have currently questioned the approach of only testing these algorithms in laboratory, as the algorithms will actually be used in free-living settings. Under free-living conditions, the activities do not occur at a pre-defined window of time (e.g., 30 s, 60 s) and are not choreographed. One study by Sasaki et al. (2016b) demonstrated that although machine learning algorithms trained with laboratory data performed almost perfectly when tested in laboratory (up to 96% accuracy), the accuracy was rather poor when scrutinized in free-living settings (< 55% accuracy). Conversely, when trained with free-living data, the algorithms performed significantly better (up to 69%), but still fell short of expectation (Sasaki et al., 2016b). Several other studies have confirmed this decrement in performance of machine learning algorithms in free-living conditions, alerting to the need of further refinements before operationalizing the use of such methods in the field.

Parameters of device use To date, most studies have adopted the waist as the placement of choice for capturing physical activity via motion sensors. This is based on the fact that energy expenditure is better reflected by the displacement of the center of mass (Sasaki et al., 2016a). The accuracy for measuring energy expenditure is a positive aspect for waist placement, there are tradeoffs with this approach, especially related to compliance of use (Troiano et al., 2014). Researchers often report difficulties in obtaining data for a significant proportion of participants or the day in research studies. Of further note, most of the times it is not feasible to request participants to use the accelerometer for 24 h, when adopting waist placement.

Physical Sensors: Motion Sensors for Physical Activity Assessment Table 2

117

ActiGraph cut-points for different age groups and populations. Sample

Activities

Equation

Cut-Points

Model Performance

SB ¼ 0–50 counts.30 s 1 LPA ¼ 51–1499 counts.30 s 1 MVPA ¼ 1500–2600 counts.30 s 1 VPA >2600 counts.30 s 1 SB ¼ 0–25 counts.15 s 1 LPA ¼ 26–573 counts.15 s 1 MVPA ¼ 574–1002 counts.15 s 1 VPA 1003 counts.15 s 1

R2 [ 0.84 SEE ¼ 1.36 MET

METs ¼ 1.439008 þ (0.000795acounts.min 1) SBa ¼ 0–99 counts.min 1 LPA ¼ 100–1951 counts.min 1 MVPA ¼ 1952–5724 counts.min 1 VPA 5725 counts.min 1 METs ¼ 2.606 þ (0.0006863acounts.min 1) SBa ¼ 0–99 counts.min 1 LPA ¼ 100–574 counts.min 1 MVPA ¼ 574–4944 counts.min 1 VPA 4945 counts.min 1 METs ¼ 0.000863(VM3 counts. SBb ¼ 0–199 counts.min 1 min 1) þ 0.668876 LPA ¼ 200–2689 counts.min 1 MVPA ¼ 2690–6166 counts.min 1 VPA 6167 counts.min 1

R2 [ 0.82 SEE ¼ 1.12 MET

Children and adolescents Treuth et al. 74 female (2004) adolescents (13–14 yo)

Locomotion and METs ¼ 2.01 þ (0.00171acounts.30 s 1) lifestyles activities

Evenson et al. (2008)

Sedentary, locomotion and playing activities

33 boys and girls (5–8 yo)

Adults Freedson et al. (1998)

25 men Locomotion (24.8  4.2 yo) and 25 women (22.9  3.8 yo) Swartz et al. 31 men (41  Locomotion and lifestyle activities (2000) 17 yo) e 39 women (42  14 yo) Sasaki et al. n ¼ 36 men and Locomotion (2011) women (27.8  8.6 yo) Older Adults Copeland and 38 older men Esliger and women (2009) (69.7  3.5 yo) Miller et al. 30 older men (2010) and women (60–69 yo)

Locomotion

Locomotion

Individuals with multiple sclerosis Motl et al. 24 individuals Locomotion (2009) with MS (20 women) (43.5  12.2 yo) Sandroff 54 persons with Locomotion et al. MS (2014) (45 women) (50.9  9.2 yo)

ROC analysis

ROC range ¼ 0.83– 0.98

R2 [ 0.32 SEE ¼ 1.16 MET

R2 ¼ 0.78 SEE ¼ 1.3 MET

Mean counts above 3.2 km h 1 (VO2 ¼ 13.0  SB ¼ 0–49 counts.min 1 – 2.1 mL kg 1 min 1) LPA ¼ 50–1040 counts.min 1 MVPA 1041 counts.min 1 %VO2 ¼ 29.238 þ (0.0055366acounts.min 1) SBa ¼ 0–99 counts.min 1 R2 ¼ 0.43 SEE ¼ 11.6% LPA ¼ 100–1907 counts.min 1 MVPA ¼ 1908–5806 counts.min 1 VPA >5807 counts.min 1 METs ¼ 2.699067 þ (.000509acounts$min 1) LPA 6460 counts.min 1

R2 ¼ 0.64 SEE ¼ 0.71 MET

Not reported

Overall Sample R2 ¼ 0.91 Mild/Moderate disability R2 ¼ 0.95/ R2 ¼ 0.94 Severe disability R2 ¼ 0.83

Overall Sample MVPA 1754 counts.min 1 Mild/Moderate disability MVPA 1980 counts.min 1 Severe disability MVPA 1185 counts.min 1

(Continued)

118 Table 2

Physical Sensors: Motion Sensors for Physical Activity Assessment ActiGraph cut-points for different age groups and populations.dcont'd Sample

Activities

Individuals with Parkinson’s disease Jeng et al. 30 individuals Locomotion (2020b) (64.4  6.4 yo)

Equation

Cut-Points

Model Performance

Not reported

LPA ¼ 0.80). The systematic review further indicated that both the Fitbit and Jawbone fitness trackers more often underestimated the energy expenditure of activities (Evenson et al., 2015). Similarly as with research-grade accelerometers, substantially fewer studies have examined the accuracy of fitness trackers for assessing physical activity in special populations. In this regard, we have recently conducted a study demonstrating that the GarminVivosmart and the Fitbit One were highly accurate in assessing steps taken during self-paced overground walking in individuals with Parkinson’s disease. The devices presented relative error values of 1.0% and 1.3%, respectively (Lai et al., 2020). Conversely, the Fitbit Charge 2 HR presented a relative error of 4.8%. Moreover, agreement levels (intra-class correlation coefficient) between step counts from these devices and manually counted steps were 0.97, 0.98 and 0.47 for the Garmin Vivosmart, Fitbit One, and Fitbit Charge 2 HR, respectively (Lai et al., 2020). The probable explanation for the worse accuracy of the Fitbit Charge 2 HR might be related to body placement, as such device was positioned on the wrist, while the other two were attached to the waist. This indicates that device placement is an issue with research or commercial grade devices. Fitness trackers present limitations and lack transparency concerning the translation of signals to physical activity metrics, but have immense potential for promoting positive behavior changes at large. Indeed, there are abundant data indicating the fitness tracker use is associated with levels of physical activity behavior in a number of populations including persons with MS (Brickwood et al., 2019; Davergne et al., 2019; Silveira et al., 2021). Therefore, researchers should continue to examine the use, applications, and validity of these devices for physical activity outcomes.

What is next? Over the next 5 years, it is expected that researchers will further test and refine the methods for predicting physical activity from accelerometry signals. We envision that, as technological advancements outpaced the development of new prediction methods, researchers will very likely explore new possibilities for the high-resolution output from current activity monitors. In addition, due to the ongoing transition from waist to wrist placement, there is a demand for new methods that can process wrist-based accelerometry and pedometry data. Cloud-based services for data storing, aggregation and processing from multicenter studies will likely increase, as such services facilitate data handling and reduce the possibility of missing data by faulty hardware. These platforms allow real-time monitoring of accelerometer use, and this may help researchers increase participant compliance with wear time and duration. By reviewing the literature, it is possible to observe that pedometers have mostly been used for obtaining recordings of total steps per day. However, it is important to mention that the use of pedometers for monitoring frequency, duration and intensity of physical activity in large-scale studies is likely to increase, as new such devices have presented improvements in memory and sensing technology, and some pedometers currently provide readings of minute-by-minute steps. Typically, the price of pedometers is much lower than accelerometers. Additionally, the development of step-rate thresholds provides researchers with an offthe-shelf method for monitoring physical activity in the field with pedometers. These facts place the pedometers as a more affordable alternative to accelerometers for researchers. As we highlighted throughout this manuscript, there is still a limited number of studies examining the use of pedometers, accelerometers, and fitness trackers in people with special conditions. In this regard, there is certainly room for developing smart systems for ongoing monitoring of physical activity data from special populations for early detection and identification of the disease progression and activity and its change over time (Sasaki et al., 2017).

120

Physical Sensors: Motion Sensors for Physical Activity Assessment

Summary We provided an overview of the use and applications of motion sensors for physical activity assessment. We demonstrated some potential of these devices and important limitations. Researchers and health professionals might take such evidence into consideration when deciding on use of pedometers and/or accelerometers in research and professional practice, and when interpreting outcomes. Future research is still necessary for continued improvement of the technology and methods associated with motion sensors in the field of physical activity assessment.

References Agiovlasitis, S., Sandroff, B.M., Motl, R.W., 2016. Step-rate cut-points for physical activity intensity in patients with multiple sclerosis: The effect of disability status. Journal of the Neurological Sciences 361, 95–100. https://doi.org/10.1016/j.jns.2015.12.027. Aguilar-Farías, N., Brown, W.J., Peeters, G.M.E.E.G., 2014. ActiGraph GT3Xþ cut-points for identifying sedentary behaviour in older adults in free-living environments. Journal of Science and Medicine in Sport / Sports Medicine Australia 17, 293–299. https://doi.org/10.1016/j.jsams.2013.07.002. Aminian, S., Hinckson, E.A., 2012. Examining the validity of the ActivPAL monitor in measuring posture and ambulatory movement in children. International Journal of Behavioral Nutrition and Physical Activity 9, 119. https://doi.org/10.1186/1479-5868-9-119. Bandura, A., 2004. Health promotion by social cognitive means. Health Education & Behavior 31, 143–164. https://doi.org/10.1177/1090198104263660. Bassett, D.R., 2012. Device-based monitoring in physical activity and public health research. Physiological Measurement 33, 1769. https://doi.org/10.1088/0967-3334/33/ 11/1769. Bassett Jr., D.R., Rowlands, A., Trost, S.G., 2012. Calibration and validation of wearable monitors. Medicine and Science in Sports and Exercise 44, S32–S38. https://doi.org/ 10.1249/MSS.0b013e3182399cf7. Bassett, D.R., Toth, L.P., LaMunion, S.R., Crouter, S.E., 2017. Step counting: A review of measurement considerations and health-related applications. Sports Medicine 47, 1303– 1315. https://doi.org/10.1007/s40279-016-0663-1. Bravata, D.M., Smith-Spangler, C., Sundaram, V., Gienger, A.L., Lin, N., Lewis, R., Stave, C.D., Olkin, I., Sirard, J.R., 2007. Using pedometers to increase physical activity and improve health: A systematic review. JAMA: The Journal of the American Medical Association 298, 2296–2304. https://doi.org/10.1001/jama.298.19.2296. Brickwood, K.-J., Watson, G., O’Brien, J., Williams, A.D., 2019. Consumer-based wearable activity trackers increase physical activity participation: Systematic review and metaanalysis. JMIR mHealth and uHealth 7, e11819. https://doi.org/10.2196/11819. Caspersen, C.J., Powell, K.E., Christenson, G.M., 1985. Physical activity, exercise, and physical fitness: Definitions and distinctions for health-related research. Public Health Reports 100, 126–131. Chen, K.Y., Bassett Jr., D.R., 2005. The technology of accelerometry-based activity monitors: Current and future. Medicine and Science in Sports and Exercise 37, S490–S500. Chen, K.Y., Janz, K.F., Zhu, W., Brychta, R.J., 2012. Redefining the roles of sensors in objective physical activity monitoring. Medicine and Science in Sports and Exercise 44, S13– S23. https://doi.org/10.1249/MSS.0b013e3182399bc8. Clanchy, K.M., Tweedy, S.M., Boyd, R.N., Trost, S.G., 2011. Validity of accelerometry in ambulatory children and adolescents with cerebral palsy. European Journal of Applied Physiology 111, 2951–2959. https://doi.org/10.1007/s00421-011-1915-2. Copeland, J.L., Esliger, D.W., 2009. Accelerometer assessment of physical activity in active, healthy older adults. Journal of Aging and Physical Activity 17, 17–30. Crouter, S.E., Schneider, P.L., Karabulut, M., Bassett, D.R., 2003. Validity of 10 electronic pedometers for measuring steps, distance, and energy cost. Medicine and Science in Sports and Exercise 35, 1455–1460. https://doi.org/10.1249/01.MSS.0000078932.61440.A2. Crouter, S.E., Schneider, P.L., Bassett, D.R., 2005. Spring-levered versus piezo-electric pedometer accuracy in overweight and obese adults. Medicine and Science in Sports and Exercise 37, 1673–1679. Crouter, S.E., Churilla, J.R., Bassett Jr., D.R., 2006. Estimating energy expenditure using accelerometers. European Journal of Applied Physiology 98, 601–612. https://doi.org/ 10.1007/s00421-006-0307-5. Crouter, S.E., Churilla, J.R., Bassett, D.R., 2008. Accuracy of the Actiheart for the assessment of energy expenditure in adults. European Journal of Clinical Nutrition 62, 704–711. https://doi.org/10.1038/sj.ejcn.1602766. Dahlgren, G., Carlsson, D., Moorhead, A., Häger-Ross, C., McDonough, S.M., 2010. Test–retest reliability of step counts with the ActivPAL™ device in common daily activities. Gait & Posture 32, 386–390. https://doi.org/10.1016/j.gaitpost.2010.06.022. Davergne, T., Pallot, A., Dechartres, A., Fautrel, B., Gossec, L., 2019. Use of wearable activity trackers to improve physical activity behavior in patients with rheumatic and musculoskeletal diseases: A systematic review and meta-analysis. Arthritis Care & Research (Hoboken) 71, 758–767. https://doi.org/10.1002/acr.23752. DeVries, S.I., Garre, F.G., Engbers, L.H., Hildebrandt, V.H., Van Buuren, S., 2011. Evaluation of neural networks to identify types of activity using accelerometers. Medicine & Science in Sports & Exercise 43, 101–107. https://doi.org/10.1249/MSS.0b013e3181e5797d. Dumbauld, E., 1946. Thomas Jefferson: American Tourist. University of Oklahoma Press. Elsworth, C., Dawes, H., Winward, C., Howells, K., Collett, J., Dennis, A., Sackley, C., Wade, D., 2009. Pedometer step counts in individuals with neurological conditions. Clinical Rehabilitation 23, 171–175. https://doi.org/10.1177/0269215508098895. Evenson, K.R., Catellier, D.J., Gill, K., Ondrak, K.S., McMurray, R.G., 2008. Calibration of two objective measures of physical activity for children. Journal of Sports Sciences 26, 1557–1565. https://doi.org/10.1080/02640410802334196. Evenson, K.R., Goto, M.M., Furberg, R.D., 2015. Systematic review of the validity and reliability of consumer-wearable activity trackers. International Journal of Behavioral Nutrition and Physical Activity 12, 159. https://doi.org/10.1186/s12966-015-0314-1. Freedson, P.S., Melanson, E., Sirard, J., 1998. Calibration of the Computer Science and Applications, Inc. accelerometer. Medicine and Science in Sports and Exercise 30, 777–781. Freedson, P.S., Lyden, K., Kozey-Keadle, S., Staudenmayer, J., 2011. Evaluation of artificial neural network algorithms for predicting METs and activity type from accelerometer data: Validation on an independent sample. Journal of Applied Physiology 111, 1804–1812. https://doi.org/10.1152/japplphysiol.00309.2011. Gibbs-Smith, C.H., Rees, G., 1978. The Inventions of Leonardo Da Vinci. Scribner. Harrington, D.M., Welk, G.J., Donnelly, A.E., 2011. Validation of MET estimates and step measurement using the ActivPAL physical activity logger. Journal of Sports Sciences 29, 627–633. https://doi.org/10.1080/02640414.2010.549499. Hendelman, D., Miller, K., Baggett, C., Debold, E., Freedson, P., 2000. Validity of accelerometry for the assessment of moderate intensity physical activity in the field. Medicine and Science in Sports and Exercise 32, S442–S449. Hickey, A., John, D., Sasaki, J.E., Mavilia, M., Freedson, P., 2015. Validity of activity monitor step detection is related to movement patterns. Journal of Physical Activity & Health. https://doi.org/10.1123/jpah.2015-0203. Holmlund, T., Ekblom-Bak, E., Franzén, E., Hultling, C., Wahman, K., 2020. Defining accelerometer cut-points for different intensity levels in motor-complete spinal cord injury. Spinal Cord 58, 116–124. https://doi.org/10.1038/s41393-019-0308-y.

Physical Sensors: Motion Sensors for Physical Activity Assessment

121

Jeng, B., Cederberg, K.L., Lai, B., Sasaki, J.E., Bamman, M.M., Motl, R.W., 2020a. Step-rate threshold for physical activity intensity in Parkinson’s disease. Acta Neurologica Scandinavica 142, 145–150. https://doi.org/10.1111/ane.13250. Jeng, B., Cederberg, K.L.J., Lai, B., Sasaki, J.E., Bamman, M.M., Motl, R.W., 2020b. Accelerometer output and its association with energy expenditure in persons with mild-tomoderate Parkinson’s disease. PLoS One 15, e0242136. https://doi.org/10.1371/journal.pone.0242136. Johannsen, D.L., Calabro, M.A., Stewart, J., Franke, W., Rood, J.C., Welk, G.J., 2010. Accuracy of armband monitors for measuring daily energy expenditure in healthy adults. Medicine and Science in Sports and Exercise 42, 2134–2140. https://doi.org/10.1249/MSS.0b013e3181e0b3ff. John, D., Miller, R., Kozey-Keadle, S., Caldwell, G., Freedson, P., 2012. Biomechanical examination of the “plateau phenomenon” in ActiGraph vertical activity counts. Physiological Measurement 33, 219–230. https://doi.org/10.1088/0967-3334/33/2/219. Kozey-Keadle, S., Libertine, A., Lyden, K., Staudenmayer, J., Freedson, P.S., 2011. Validation of wearable monitors for assessing sedentary behavior. Medicine and Science in Sports and Exercise 43, 1561–1567. https://doi.org/10.1249/MSS.0b013e31820ce174. Lai, B., Sasaki, J.E., Jeng, B., Cederberg, K.L., Bamman, M.M., Motl, R.W., 2020. Accuracy and precision of three consumer-grade motion sensors during Overground and treadmill walking in people with Parkinson disease: Cross-sectional comparative study. JMIR Rehabilitation and Assistive Technologies 7, e14059. https://doi.org/10.2196/14059. Lyden, K., Keadle, S.K., Staudenmayer, J., Freedson, P.S., 2017. The activPAL™ accurately classifies activity intensity categories in healthy adults. Medicine and Science in Sports and Exercise 49, 1022–1028. https://doi.org/10.1249/MSS.0000000000001177. Maddocks, M., Petrou, A., Skipper, L., Wilcock, A., 2010. Validity of three accelerometers during treadmill walking and motor vehicle travel. British Journal of Sports Medicine 44, 606–608. https://doi.org/10.1136/bjsm.2008.051128. Mannini, A., Intille, S.S., Rosenberger, M., Sabatini, A.M., Haskell, W., 2013. Activity recognition using a single accelerometer placed at the wrist or ankle. Medicine and Science in Sports and Exercise 45, 2193–2203. https://doi.org/10.1249/MSS.0b013e31829736d6. Matthews, C.E., 2005. Calibration of accelerometer output for adults. Medicine & Science in Sports & Exercise 37 (Suppl.), S512–S522. https://doi.org/10.1249/01.mss. 0000185659.11982.3d, 2005. Matthews, C.E., Chen, K.Y., Freedson, P.S., Buchowski, M.S., Beech, B.M., Pate, R.R., Troiano, R.P., 2008. Amount of time spent in sedentary behaviors in the United States, 2003–2004. American Journal of Epidemiology 167, 875–881. https://doi.org/10.1093/aje/kwm390. Miller, N.E., Strath, S.J., Swartz, A.M., Cashin, S.E., 2010. Estimating absolute and relative physical activity intensity across age via accelerometry in adults. Journal of Aging and Physical Activity 18, 158–170. Montoye, H.J., Washburn, R., Servais, S., Ertl, A., Webster, J.G., Nagle, F.J., 1983. Estimation of energy expenditure by a portable accelerometer. Medicine and Science in Sports and Exercise 15, 403–407. Motl, R.W., McAuley, E., Snook, E.M., Scott, J.A., 2005. Accuracy of two electronic pedometers for measuring steps taken under controlled conditions among ambulatory individuals with multiple sclerosis. Multiple Sclerosis Journal 11, 343–345. https://doi.org/10.1191/1352458505ms1161oa. Motl, R.W., Snook, E.M., Agiovlasitis, S., Suh, Y., 2009. Calibration of accelerometer output for ambulatory adults with multiple sclerosis. Archives of Physical Medicine and Rehabilitation 90, 1778–1784. https://doi.org/10.1016/j.apmr.2009.03.020. O’Brien, M.W., Kivell, M.J., Wojcik, W.R., D’Entremont, G.R., Kimmerly, D.S., Fowles, J.R., 2018. Influence of anthropometrics on step-rate thresholds for moderate and vigorous physical activity in older adults: Scientific modeling study. JMIR Aging 1, e12363. https://doi.org/10.2196/12363. Romanzini, M., Petroski, E.L., Reichert, F.F., 2012. Accelerometers thresholds to estimate physical activity intensity in children and adolescents: A systematic review. Revista Brasileira de Cineantropometria & Desempenho Humano 14, 101–113. https://doi.org/10.5007/1980-0037.2012v14n1p101. Rosenberger, M.E., Fulton, J.E., Buman, M.P., Troiano, R.P., Grandner, M.A., Buchner, D.M., Haskell, W.L., 2019. The 24-hour activity cycle: A new paradigm for physical activity. Medicine and Science in Sports and Exercise 51, 454–464. https://doi.org/10.1249/MSS.0000000000001811. Sandroff, B.M., Motl, R.W., Suh, Y., 2012. Accelerometer output and its association with energy expenditure in persons with multiple sclerosis. Journal of Rehabilitation Research and Development 49, 467–475. Sandroff, B.M., Klaren, R.E., Pilutti, L.A., Motl, R.W., 2014. Oxygen cost of walking in persons with multiple sclerosis: Disability matters, but why? Multiple Sclerosis International 2014, 162765. https://doi.org/10.1155/2014/162765. Sasaki, J.E., John, D., Freedson, P.S., 2011. Validation and comparison of ActiGraph activity monitors. Journal of Science and Medicine in Sport 14, 411–416. https://doi.org/ 10.1016/j.jsams.2011.04.003. Sasaki, J.E., Hickey, A., Mavilia, M., Tedesco, J., John, D., Kozey Keadle, S., Freedson, P.S., 2015. Validation of the fitbit wireless activity tracker for prediction of energy expenditure. Journal of Physical Activity & Health 12, 149–154. https://doi.org/10.1123/jpah.2012-0495. Sasaki, J.E., Da Silva, K.S., Da Costa, B.G.G., John, D., 2016a. Measurement of physical activity using accelerometers. In: Computer-Assisted and Web-Based Innovations in Psychology, Special Education, and Health. Academic Press, pp. 33–60. Sasaki, J.E., Hickey, A.M., Staudenmayer, J.W., John, D., Kent, J.A., Freedson, P.S., 2016b. Performance of activity classification algorithms in free-living older adults. Medicine and Science in Sports and Exercise 48, 941–950. https://doi.org/10.1249/MSS.0000000000000844. Sasaki, J.E., Sandroff, B., Bamman, M., Motl, R.W., 2017. Motion sensors in multiple sclerosis: Narrative review and update of applications. Expert Review of Medical Devices 14, 891–900. https://doi.org/10.1080/17434440.2017.1386550. Sasaki, J.E., Júnior, J.H., Meneguci, J., Tribess, S., Marocolo Júnior, M., Stabelini Neto, A., Virtuoso Júnior, J.S., 2018. Number of days required for reliably estimating physical activity and sedentary behaviour from accelerometer data in older adults. Journal of Sports Sciences 36, 1572–1577. https://doi.org/10.1080/02640414.2017.1403527. Schneider, P.L., Crouter, S.E., Lukajic, O., Bassett, D.R., 2003. Accuracy and reliability of 10 pedometers for measuring steps over a 400-m walk. Medicine and Science in Sports and Exercise 35, 1779–1784. https://doi.org/10.1249/01.MSS.0000089342.96098.C4. Silveira, S.L., Baird, J.F., Motl, R.W., 2021. Rates, patterns, and correlates of fitness tracker use among older adults with multiple sclerosis. Disability and Health Journal 14, 100966. https://doi.org/10.1016/j.dhjo.2020.100966. Staudenmayer, J., Pober, D., Crouter, S., Bassett, D., Freedson, P., 2009. An artificial neural network to estimate physical activity energy expenditure and identify physical activity type from an accelerometer. Journal of Applied Physiology 107, 1300–1307. https://doi.org/10.1152/japplphysiol.00465.2009. Staudenmayer, J., He, S., Hickey, A., Sasaki, J., Freedson, P., 2015. Methods to estimate aspects of physical activity and sedentary behavior from high-frequency wrist accelerometer measurements. Journal of Applied Physiology 119, 396–403. https://doi.org/10.1152/japplphysiol.00026.2015. Swartz, A.M., Strath, S.J., Bassett Jr., D.R., O’Brien, W.L., King, G.A., Ainsworth, B.E., 2000. Estimation of energy expenditure using CSA accelerometers at hip and wrist sites. Medicine and Science in Sports and Exercise 32, S450–S456. Treuth, M.S., Schmitz, K., Catellier, D.J., McMurray, R.G., Murray, D.M., Almeida, M.J., Going, S., Norman, J.E., Pate, R., 2004. Defining accelerometer thresholds for activity intensities in adolescent girls. Medicine and Science in Sports and Exercise 36, 1259–1266. Troiano, R.P., McClain, J.J., Brychta, R.J., Chen, K.Y., 2014. Evolution of accelerometer methods for physical activity research. British Journal of Sports Medicine 48, 1019–1023. https://doi.org/10.1136/bjsports-2014-093546. Trost, S.G., McIver, K.L., Pate, R.R., 2005. Conducting accelerometer-based activity assessments in field-based research. Medicine and Science in Sports and Exercise 37, S531–S543. Trost, S.G., Wong, W.-K., Pfeiffer, K.A., Zheng, Y., 2012. Artificial neural networks to predict activity type and energy expenditure in youth. Medicine and Science in Sports and Exercise 44, 1801–1809. https://doi.org/10.1249/MSS.0b013e318258ac11. Tudor-Locke, C., Bassett, D.R., 2004. How many steps/day are enough? Sports Medicine 34, 1–8. https://doi.org/10.2165/00007256-200434010-00001.

122

Physical Sensors: Motion Sensors for Physical Activity Assessment

Tudor-Locke, C., Craig, C.L., Aoyagi, Y., Bell, R.C., Croteau, K.A., De Bourdeaudhuij, I., Ewald, B., Gardner, A.W., Hatano, Y., Lutes, L.D., Matsudo, S.M., Ramirez-Marrero, F.A., Rogers, L.Q., Rowe, D.A., Schmidt, M.D., Tully, M.A., Blair, S.N., 2011a. How many steps/day are enough? For older adults and special populations. International Journal of Behavioral Nutrition and Physical Activity 8, 80. https://doi.org/10.1186/1479-5868-8-80. Tudor-Locke, C., Craig, C.L., Beets, M.W., Belton, S., Cardon, G.M., Duncan, S., Hatano, Y., Lubans, D.R., Olds, T.S., Raustorp, A., Rowe, D.A., Spence, J.C., Tanaka, S., Blair, S.N., 2011b. How many steps/day are enough? For children and adolescents. International Journal of Behavioral Nutrition and Physical Activity 8, 78. https://doi.org/ 10.1186/1479-5868-8-78. Tudor-Locke, C., Craig, C.L., Brown, W.J., Clemes, S.A., De Cocker, K., Giles-Corti, B., Hatano, Y., Inoue, S., Matsudo, S.M., Mutrie, N., Oppert, J.-M., Rowe, D.A., Schmidt, M.D., Schofield, G.M., Spence, J.C., Teixeira, P.J., Tully, M.A., Blair, S.N., 2011c. How many steps/day are enough? For adults. International Journal of Behavioral Nutrition and Physical Activity 8, 79. https://doi.org/10.1186/1479-5868-8-79. Tudor-Locke, C., Williams, J.E., Reis, J.P., Pluto, D., 2012. Utility of pedometers for assessing physical activity. Sports Medicine 32, 795–808. https://doi.org/10.2165/00007256200232120-00004. Tudor-Locke, C., Aguiar, E.J., Han, H., Ducharme, S.W., Schuna, J.M., Barreira, T.V., Moore, C.C., Busa, M.A., Lim, J., Sirard, J.R., Chipkin, S.R., Staudenmayer, J., 2019. Walking cadence (steps/min) and intensity in 21–40 year olds: CADENCE-adults. International Journal of Behavioral Nutrition and Physical Activity 16, 8. https://doi.org/10.1186/s12966019-0769-6. Warburton, D.E.R., Nicol, C.W., Bredin, S.S.D., 2006. Health benefits of physical activity: The evidence. CMAJ 174, 801–809. https://doi.org/10.1503/cmaj.051351. Zhang, S., Murray, P., Zillmer, R., Eston, R.G., Catt, M., Rowlands, A.V., 2012a. Activity classification using the GENEA: Optimum sampling frequency and number of axes. Medicine and Science in Sports and Exercise 44, 2228–2234. https://doi.org/10.1249/MSS.0b013e31825e19fd. Zhang, S., Rowlands, A.V., Murray, P., Hurst, T.L., 2012b. Physical activity classification using the GENEA wrist-worn accelerometer. Medicine and Science in Sports and Exercise 44, 742–748. https://doi.org/10.1249/MSS.0b013e31823bf95c.

Physical Sensors: Holographic Sensors Christopher R. Lowea and Gita Khalili Moghaddamb, a Cambridge Academy of Therapeutic Sciences (CATS), University of Cambridge, Cambridge, United Kingdom; and b Department of Clinical Neurosciences, Cambridge Biomedical Campus, Addenbrooke’s Hospital, Cambridge, United Kingdom © 2023 Elsevier Ltd. All rights reserved.

Introduction Holography and holographic transduction Pioneering holographic sensors Bespoke holographic ion sensors Holographic metabolite sensors Holographic glucose sensors The lactate challenge Holographic sensors for microorganisms Holographic sensors for physical parameters New holographic procedures Chemical and combinatorial holography Miniaturization of holographic sensors Modelling and quantitation of holographic sensors Conclusions References

123 124 125 127 128 130 131 132 133 133 135 135 136 138 139

Abstract Holographic sensors are an exciting and versatile new addition to the analysts armamentarium which offer a number of advantages over conventional sensor technologies: They are an inexpensive and scalable platform amenable to mass production, configurable in multiple formats, including miniaturization, arrays and free standing flakes, and are able to output the results qualitatively or quantitatively as changes in color, intensity, two- and three-dimensional images and alphanumeric displays and can be assessed visually and using smartphones. This review describes the basics of holography, how relatively simple holographic sensors were first developed and how the field has developed such that a plethora of key analytes including ions, gases, volatiles, metabolites, drugs, proteins, enzymes, inhibitors and whole cells could be monitored with holograms constructed with novel bespoke hydrophilic and apolar matrices and suitable embedded recognitionresponse systems. More recent developments such as molecular imprinting, combinatorial chemistry, increased reflectivity, new photopolymer materials, micro array technologies and improved modelling greatly enhance their selectivity, sensitivity and general acceptability. Nevertheless, despite a requirement for further research, this holographic sensor platform is highly versatile and can be configured to create visual or instrumentally monitored responses in discreet or real-time at macro, micro and nano scales.

Introduction It has been a long term aspiration that sensors exploiting highly selective recognition elements coupled to appropriate physicochemical transducers would displace most other analytical technologies within the healthcare, consumer, food and beverage, sports, lifestyle and recreation, environmental, defense and security, biomedical and life science industries by giving direct readouts to the user in near real time. To some extent, after nearly 4–5 decades of research, sensor technologies are coming of age, with the principal driver being the growing inclination towards automation, especially in the Asia-Pacific region. Estimates of market size in 2024–25 are in the range $287–323B (www.alliedmarketresearch.com; www.bccresearch.com) with CAGRs in the range 7–9.5%. However, these sensor applications are largely based on the adoption of robust physical sensors for monitoring parameters such as motion, strain, acceleration, light, sound, temperature, pressure and magnetic field intensity. The development, introduction and up-take of sensors which respond specifically to chemicals or biological agents in various analytical scenarios is a long way behind physical sensors and has, to a large extent, evaded translation into reality. The original appeal of biosensors, i.e. their low cost, small size, fast and real-time response, ready interface with computers and smart phones, and facile use by lay personnel in a userfriendly fashion, has yet to be realized. This is primarily because early embodiments of biosensors lacked durability, were expensive and were unsuited to manufacture at scale. More recently, these issues are being addressed by coupling recognition systems to novel transducer technologies that have been developed in the micro- and plastic-electronics, printing, paper and photographic industries and are thus more amenable to mass production. Our approach is to use a reflection hologram embedded in a “smart” hydrogel as the interactive element in an inexpensive, user friendly and truly mass producible sensor platform. This concept is unique in the

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.10037-8

123

124

Physical Sensors: Holographic Sensors

field of chemical and biological sensors in that the reflection hologram itself provides both the analyte-selective matrix and the optical interrogation and reporting transducer combined in a variety of operational configurations (Lowe et al., 1995; Blyth et al., 1996; Lowe, 1999, 2007; Lowe and Larbey, 2008).

Holography and holographic transduction Holography was conceived by the Hungarian physicist Dennis Gabor and published in Nature in 1948 (Gabor, 1948). Appropriately, the word hologram is derived from the Greek words ólo2 (holos) meaning “whole,” and grammé (gramma), meaning “message.” Gabor’s discovery was serendipitous, made while trying to mitigate spherical aberration in the lenses to improve the resolution of electron microscopes, and did not have any other known applications at the time. Indeed, in September 1948 the New York Times described the hologram as “a futuristic tapestry that mysteriously recreated images out of thin air,” and until the late 1950s the technique was deemed “unintuitive and baffling,” of “dubious practicality” and even pronounced a “white elephant”! The discovery of holography was an unexpected result of research into improving electron microscopes at the British Thomson-Houston Company in Rugby, although further development in the field was stymied during the next decade or so because light sources available at the time were not mechanically or thermally stable or truly “coherent” and had to await the discovery of the emission of monochromatic light by the laser in 1960. The first true holograms which recorded 3D objects were fabricated independently by Emmett Leith and Juris Upatnieks in the USA and by Yuri Denisyuk in the former Soviet Union in 1962 (Leith and Upatnieks, 1962; Denisyuk, 1962). The original transmission holograms of Leith and Upatnieks (1962) were fabricated with silver halide chemistry and required monochromatic light to view the image to avoid chromatic aberrations, while those from Denisyuk (1962) were based on the color photography pioneered by Lippmann (1891, 1894) and could be viewed in broad spectrum light. Nowadays, holograms produced by Denisyuk’s methodology are well established with applications in displays, data storage, metamaterials, 3D imaging, non-destructive testing, interferometry, optical physics and telecommunications. In a further refinement, Stephen Benton (1972), working for Polaroid Inc., patented a method for circumventing chromatic aberration in transmission holograms which subsequently formed the basis for embossable holograms, so-called Benton or Rainbow holograms, since they diffract a rainbow of colors on illumination with white light. These holograms are the basis for > 99% of the current holographic industry market, particularly security holograms on credit cards, tickets and bank notes, and as decorative foils and wrappings. In their simplest form (Fig. 1), holograms can be fabricated by passing a single laser beam through a beam splitter, which divides the beam into two beams: The first beam is expanded by a lens, and redirected by front surface mirrors onto an object. The light that is scattered back falls onto a recording medium. Meanwhile, the second beam, expanded by a lens, travels directly onto the recording medium. The interference of the two mutually coherent beams creates constructive (antinodes) and destructive (nodes) interferences, which are recorded in the photosensitive medium as interference fringes, and thus all optical information about the object is encoded in the diffraction field produced by the hologram at the reconstruction stage (Hariharan, 2010). The contrast between the node and antinode fringes is determined by the coherence length of the laser light, the state of polarization of the two recording beams, and the ratio of their intensities. The coherence length is defined as the distance L over which a wave maintains a constant phase relation or coherence, and is defined as L ¼ l02/2Dl, where.

Fig. 1

General principle of holography.

Physical Sensors: Holographic Sensors

125

l0 is the mean wavelength (l0  Dl) and Dl the bandwidth of the radiation. High contrast interference fringes are essential to achieve effective holograms and require that the optical path difference between the object and reference beams be small compared to the coherence length of the radiation (Born and Wolf, 1999). The visibility (V) of the fringes with an optical path difference less than the coherence length (L) is determined by three factors: the spatial coherence (m), the angle between the planes of polarization (4) and the ratio (R) of the intensities of the object (Io) and reference (Ir) waves, i.e. R¼ Io/Ir (Collier et al., 1971). The visibility (V) of the interference fringes is given by: V ¼ mð2OIr Io =Ir þ Io Þ ¼ 2mðOR=1 þ RÞ cos 4 whence the maximum visibility or contrast of the fringes (V) of 1 is realised when R ¼ 1, i.e. the intensities of Io and Ir are equal and the two beams have identical planes of polarizations (cos 4 ¼ 1). One of the simplest approaches to recording holograms is in Denisyuk reflection mode in which fringes are recorded in the optical properties of a recording material. In this mode, the reference (Ir) to object (Io) beam ratio (R) is typically 1.33 (Saxby, 2004). Such reflection holograms are usually created by passing an expanded beam of laser light through the recording plate to illuminate an object on the other side of the plate, whence light reflected from the object (Io) interferes with the illuminating light (Ir) thereby forming standing waves which are recorded as periodic modulations of refractive index, i.e. fringes or a grating, running approximately parallel to the plane of the recording material. When the hologram is illuminated with a white light source, the recorded fringe spacing act as Bragg mirrors and the diffracted light forms an image of the original object used during exposure to the illuminating laser (Saxby, 2004; Benton and Bove, 2007). The diffracted light from the periodic gratings results in a narrow-band spectral peak determined by the wavelength of the incident laser light and the angle between the two recording beams. The holographic diffraction is governed by Bragg’s law (Fig. 2): lmax ¼ 2n0 vsinq where lmax is the wavelength of the first order diffracted light, n0 is the effective refractive index of the recording matrix, v is the spacing between the fringes and q is the Bragg angle determined by the recording geometry. Single beam reflection Denisyuk holograms allow the image to be viewed by the observer from the same side as the hologram was originally illuminated and thus form an ideal basis for a deceptively simple, generic approach for the fabrication of chemical sensors incorporating reflection holograms embedded within “smart” hydrogel films (Lowe et al., 1995; Blyth et al., 1996; Lowe, 1999; Yetisen et al., 2014a). Typically, Denisyuk gratings comprise of thin silver halide-gelatine photographic emulsions coated onto glass or plastic substrates and are fabricated by passing a single collimated laser beam through the film backed by a front surface silvered mirror. Interference between the incident (Ir) and reflected (Io) beams creates a standing wave interference pattern, which after suitable development and fixing, creates a three-dimensional pattern of ultra-fine grains of metallic silver (Ag0;  20 nm diameter) distributed as parallel fringes separated by a distance of approximately half the wavelength (½l) of the laser light used in their construction within the thickness ( 5-10 mm) of the gelatine film (Fig. 2).

Pioneering holographic sensors The earliest example of what may be regarded as a holographic sensor arose out of an issue experienced with those making reflection or “Denisyuk” holograms. It was observed that the reconstructed image generated by illumination with a point white light source appeared at a shorter wavelength than the original laser used to fabricate it (Vilkomerson and Bostwick, 1966; Bryngdahl, 1972). For example, if the hologram was created with a HeNe laser (632.8 nm), following processing the Bragg grating compresses and the reconstructed image typically appeared green ( 532 nm). The issue was resolved by controlling the thickness of the emulsion via pre-swelling or pre-shrinking prior to laser exposure through the use of solvents (water; propan-2-ol) or proteolytic enzymes (collagenase, maxatase, trypsin and papain) (Benton and Walker, 1991; Spooncer et al., 1992). These innovations led to the first true holographic sensors aimed at measuring water in solvents (Blyth et al., 1996). When a HeNe laser-induced gelatine-based holographic grating was immersed in a test fluid, absorption of water caused the gelatine to swell perpendicular to the glass substrate, thereby increasing the fringe spacing (v) and selecting longer wavelengths to be selected for reflection from the holographic mirror. For example, the diffraction color changed as a function of the water activity (aw) when immersed in a “wet” hydrophobic solvent. The holographic sensor was able to measure the water contents of hydrocarbon solvents at sensitivities comparable to that of the gold-standard Karl Fischer coulometric titrator over a wide range of water contents. A sensor immersed in xylene recorded a visible color change when the water content was increased from 47 to 12 ppm. Since gelatine is a protein its use as the supporting matrix for a holographic grating sensor for monitoring the activity of proteolytic enzymes has also been demonstrated (Millington et al., 1995). A range of trypsin concentrations down to 25 nM within 20 min resulted in a red-shift change in color (wavelength) or brightness when the enzyme cleaves at peptide bonds adjacent to arginine and lysine residues and causes the hologram to swell. This work demonstrated the concept of a low cost, reagentless and instrumentfree sensor for trypsin with a sensitivity below that of normal duodenal levels. This approach has been improved and extended to obtain specific responses to other hydrolytic enzymes by creating holographic gratings in non-gelatine matrices which contain the substrate for the enzyme being analyzed. For example, the concentration of the digestive enzyme a-amylase was measured from the rate of reduction of diffraction efficiency at a certain wavelength over 15 min of a potato starch based hologram (Blyth et al., 1999a, b). This study demonstrated that as the enzyme degraded the starch, the peak wavelength collapsed and was replaced by a second

126

Physical Sensors: Holographic Sensors

Fig. 2 Typical set-up for the fabrication of silver nanoparticle holographic sensors showing (A) the optical bench layout, (B) the 4 offset angle to differentiate the diffracted object beam (lmax) from the specular reflection, a micrograph showing the holographic Ag0 fringes, a typical reflection spectrum (C) Replay colors ranging from the UV to the near IR part of the spectrum, (D) potential analytes and holographic characteristics which are amenable to modulation. Part of this figure (A) is reproduced with permission from Yetisen AK, Qasim MM, Nosheen S, Wilkinson TD, and Lowe CR (2014) Journal of Materials Chemistry C 2: 3569–3576.

peak at a shorter wavelength with a Bragg peak shift of 30 nm in 30 min. The effect was specific for a-amylase, since neither b-amylase nor maltase showed any effect on holograms constructed in potato starch, although the mechanism underpinning the wavelength change is not well understood. Since both gelatine and potato starch are heterogeneous polymers their use in highly selective chemical sensors is somewhat limited. This lead to the development of a new silver halide-containing holographic recording material based on a fine grain silver bromide emulsion suspended in a poly(vinyl alcohol) (PVA) matrix crosslinked with CrIII ions (Mayes et al., 1998). It was deemed

Physical Sensors: Holographic Sensors

127

that crosslinking would impart adequate structural and spatial integrity to allow a holographic image to be recorded, while permitting sufficient porosity and elasticity of the supporting polymer matrix for sensing applications. A rationally designed holographic sensor for trypsin was developed by incorporating a trypsin substrate, poly(L-lysine), into poly(vinyl alcohol) holograms to create the first “designed” holographic material which was degraded by trypsin in a concentration-dependent manner. This system was not as sensitive as that using gelatine as the holographic support matrix (Millington et al., 1995, 1996), although the degradable material, poly(L-lysine), in the PVA hologram represented only about 5% of the dry mass of the supporting matrix, compared to 100% for the gelatine, and attempts to increase the loading of the poly(L-lysine) above this level caused catastrophic grain growth in the holographic film and prevented holograms from being recorded. The use of non-gelatine based supporting polymers was further extended into fabricating diffraction gratings into poly(hydroxyethyl)methacrylate (pHEMA) films for monitoring the concentration of ethanol in samples (Blyth et al., 1999a, b; Mayes et al., 1999). The performance of the holographic sensor was demonstrated by measuring the alcohol contents of a range of alcoholic beverages, including beers, wines and spirits, without sample pre-treatment. The measured concentrations of ethanol in the beverages were determined to be within  0.3 vol % of their stated values by observing the concentration-dependent red shift in wavelength. It was also shown that to be effective in use and storage, the sensor was highly stable, unaffected by highly colored and turbid samples, and relatively insensitive to pH in the range 3–6.5. Functional modifications to this base matrix also allowed development of pH and trypsin-sensitive holograms (Mayes et al., 1999; Marshall et al., 2003). This study demonstrated not only the value of pHEMA as a support matrix for holographic sensors, but also, highlighted the need to record reflection holograms in non-conventional bespoke materials by transforming pre-formed polymer films into holographic recording media. A double-diffusion technique was devised to allow the formation of light-sensitive silver halide crystals in the pHEMA matrix (Blyth et al., 1999a, b). Typically, the holographic recording material is created by uniformly coating a thin layer ( 7–10 mm thick) of unsensitized polymer on a silanised glass slide and then immersing the slide sequentially in solutions of a silver salt (AgClO4) and a bromide salt (LiBr) containing a photosensitizing dye (1,10 -diethyl-2,20 -cyanine iodide) to form ultra-fine grains of photosensitive silver bromide (AgBr, < 20 nm dia) within the matrix of the film. Holograms were typically recorded in these films with a single 10 ns pulse from a frequency doubled Nd:YAG laser (532 nm). The photosensitive films were positioned over a front surface silvered mirror held at an angle of  4 relative to its surface in order that the finished hologram replays diffracted light at an angle differentiable from specularly reflected light. After conventional photographic development, an interference pattern comprising ultra-fine grains of metallic silver (Ag0) spaced ½l apart are generated within the thickness of the polymer film. Illumination of the grating under white light recreates the monochromatic image of the plane mirror used in its construction with the constructive interference at each fringe plane, resulting in a characteristic spectral peak with a wavelength governed by the Bragg equation and with similar brightness to those made from commercially available holographic recording materials (Blyth et al., 1996). This double-diffusion methodology circumvented the traditional laborious, emulsion-forming, techniques that were the mainstay for constructing holographic recording materials (Leubner et al., 1980). Significantly, for the fabrication of sensor holograms, the double diffusion method makes it possible to record silver halide volume holograms in a wide range of previously inaccessible natural and synthetic polymer films, even those that are somewhat hydrophobic (Mayes et al., 1999) or would otherwise encourage very rapid grain-growth of the colloidal silver particles (Lowe et al., 2003a, b). The finished holograms are robust and relatively unresponsive to changes in the physical environment. For example, the holograms are not light sensitive, since the grating consists only of Ag0 metal and, with a suitable choice of polymeric base matrix, show limited or no change in color with temperature, although it is advisable to thermostat the temperature of the medium in which the analysis is performed to  1  C. The simplicity and rapidity of this technique, combined with the sensing capability of “smart” hydrogels, offers the possibility to construct a range of inexpensive optical sensors selective for many putative analytes. “Smart” hydrogels comprise 3D polymer networks characterized by their capacity to absorb water to varying extents, from mildly absorbing by retaining  30% (w/w) in their structure, to strongly absorbing where swelling can result in a 1000% change in their original volume (Seidel and Malmonge, 2000). The hydrophilic nature of the base polymer chains encourages hydration, although this is restricted by covalent cross-linking of the polymer chains or by secondary forces such as ionic interactions, hydrogen bonding or hydrophobic forces (Flory, 1953).

Bespoke holographic ion sensors Hydrogels may be tailored to undergo macroscopic volumetric changes in response to relatively small changes in environmental conditions induced by numerous physical and chemical stimuli they can form the basis of analyte-sensitive holographic sensors which respond to a variety of putative analytes. The measurement of the concentration of protons (Hþ) and many ionic species is of paramount importance in the biomedical, environmental, food, beverage, security, agricultural and biotechnology industries. Bragg gratings can be recorded in hydrogel matrices containing functional acidic or basic monomers co-polymerized within the polymer backbone (Marshall et al., 2003). The ionization of the pendant functions causes the holographic grating to swell or contract as a result of electrostatic or osmotic forces that exchange counter-ions and water between the gel and bulk phases, changes the fringe separation (v) and selects longer or shorter wavelengths for reflection from the planar holographic mirror. The equilibrium degree of swelling of pH-sensitive hydrogels is primarily influenced not only by the charge on the ionic monomer, the pKa of the ionizable group, the relative concentration in the polymer (mol%), the density of crosslinks (mol%) and the relative polarity/

128

Physical Sensors: Holographic Sensors

apolarity of the polymer backbone, but also by the characteristics of the swelling solution, including the pH, ionic strength and nature and valence of the counter ion. The effect of pH on the replay wavelength (lmax) of a series of pHEMA holograms containing 5 mol% ethylene glycol dimethacrylate as crosslinker and 6 mol% of the functional monomers, methacrylic acid (MAA), trifluoromethylpropenoic acid (TFMPA), dimethylaminoethyl methacrylate (DMAEM), and vinyl imidazole (VI), showed classical sigmoidal titration curves, with Dlmax values up to 350 nm over their effective range (pKa  1) (Marshall et al., 2003). Curve fitting to a standard pKa program yielded pKa values of 4.56 (VI), 4.45 (TFMPA), 6.01 (MAA), and 6.93 (DMAEM) at 6 mol%. All of the matrix-bound pendant groups had their apparent pKa values shifted by up to 1.51 pH units, with the anionic groups (VI, DMAEM) downshifted and the acidic groups (TFMPA, MAA) upshifted. These shifts in the apparent pKa values of the matrix bound functional groups reflect the fact that the dissociation of these groups is influenced by their near neighbors in the polymeric environment and the hydrophobic character of the pHEMA supporting polymer. Substituting a more hydrophilic support polymer into the system alters the pH behavior of the hologram. For example, holographic sensors constructed in co-polymers of acrylamide and methacrylamide (2:1 mol%) crosslinked with N,N0 -methylenebis(acrylamide) (5 mol%) containing increasing amounts of acrylic acid (AA) (0-12 mol%) showed similar pH responses to the pHEMA holograms. Nevertheless, the apparent pKa of the matrix-bound acidic group was 4.58 (6 mol%) compared to 4.22 in the monomer, with the difference (DpKa) (0.36) being smaller than that observed in the pHEMA-based gels (1.51). This can be attributed to the fact that the pHEMA system is more hydrophobic than the acrylamide system, and polymer hydrophobicity has been shown previously to decrease the strength (i.e., the pKa) of acidic or basic comonomers (Wen et al., 1991). The effect of crosslinker concentration, buffer composition and ionic strength, temperature and reversibility of the pH-sensitive holograms was also assessed (Marshall et al., 2003). The response times of pHEMA-co-methacrylic acid pH sensors to reach a new equilibrium (lmax) following step pH changes within the range 4–6 were proportional to the concentration of ionizable monomer (mol%), the magnitude of the pH step and were inversely proportional to the buffer composition. Nevertheless, under specified conditions, the Dlmax changes are freely reversible and directly proportional to the pH values in both upward and downward directions and with no sign of hysteresis in the holographic response. As proof of utility, a pHEMA-co-methacrylic acid holographic sensor was able to monitor the changes in pH in real time in a sample of milk undergoing homolactic fermentation with Lactobacillus casei (Marshall et al., 2003). Holographic sensors for monitoring ionic strength were fabricated with a positively charged monomer, 3-(acrylamidopropyl) trimethyl ammonium chloride (ATMA, 20 mol%), a negatively charged monomer, 2-acrylamido-2-methyl-1-propanesulphonic acid (AMPS, 20 mol%) or both (ATMA (10 mol%); AMPS (10 mol%)) and crosslinked with N,N0 -methylene-bis-acrylamide (MBA, 5 mol%) (Marshall et al., 2004a). All of the holograms replayed at a wavelength of  292 nm in deionized water, which is consistent with the exposure wavelength of 532 nm and were tested with increasing concentration of NaCl up to 500 mM. The diffraction peak (lmax) of a polyacrylamide-co-AMPS-co-ATMA hologram red shifted by 58 nm from its initial value of 492 nm to a final value of 550 nm. Unlike holographic sensors that employ monomers of a single polarity, these ionic strength sensors were able to quantify ionic strength independent of the nature of ionic species in the test sample and were able to quantify the ionic strength of milk solutions containing a complex melange of ions and biological components (Marshall et al., 2004a). The rational modification of a pHEMA support matrix with crown ethers allowed the development of holographic sensors for specific monovalent cations such as Kþ and Naþ (Mayes et al., 2002) (Fig. 3). Methacrylated crown ethers were co-polymerized into a pHEMA film and the resulting holograms containing 0–97 mol% 12-crown-4, 15-crown-5 and 18-crown-6 were shown to respond to alkali and alkaline earth ions with varying magnitudes (Dlmax) and specificity. Optimized crown ether compositions containing 50 mol% receptor were able to elicit Dlmax responses of  200 nm within 30s at ion concentrations  30 mM. Potassium ion-specific holograms containing 18-crown-6 responded linearly over physiological ranges of Kþ, even in the presence of varying background levels of Naþ ( 130–150 mM) and demonstrating their potential for measuring Kþ levels in blood (Mayes et al., 2002). Holographic sensors for the real-time detection of physiological (Ca2þ, Mg2þ) and heavy (Ni2þ, Co2þ, Zn2þ) metal ions have been created by incorporating a methacrylated analog of iminodiacetic acid (IDA) into a pHEMA matrix crosslinked with ethylene glycol dimethacrylate (EDMA) (Madrigal-González et al., 2005). The sensors have been assessed for the effects of active monomer, cross-linker, pH and ionic strength on the swelling of the matrix and on metal detection sensitivity. Sensors containing > 10 mol% of chelating monomer and 6 mol% of cross-linker showed significant responses ( 46 nm) within 30s at ion concentrations of 0– 40 mM and displayed a selectivity towards the different ions tested of Ni2þ > Zn2þ > Co2þ > Ca2þ > Mg2þ. The sensor showed fully reversible responses, permitting real-time monitoring of calcium ion efflux during the germination of Bacillus megaterium spores (Madrigal-González et al., 2005).

Holographic metabolite sensors Holographic sensors are also able to detect the products of enzymatic reactions; the concept of utilizing enzyme-linked holographic sensors has been demonstrated for the clinically relevant metabolite urea and the antibiotic penicillin G (Marshall et al., 2004b, 2006). The action of the enzymes urease (EC 3.5.1.5) and penicillinase (EC 3.5.2.6) on their respective substrates, urea and penicillin G, is known to cause acidification or alkalization of test solutions. In the case of the urease-coupled holographic sensor for urea, HEMA, DMAEM and EDMA were copolymerized and the resulting hologram exposed to a urease solution (100 U/mg;

Physical Sensors: Holographic Sensors

129

Fig. 3 Holographic sensor for Kþ based on 50 mol% Methacryloylated 18-crown-6 (A) in a polyHEMA matrix showing (B) the Dlmax response to Naþ and Kþ and (C) the visual response to 20 mM Kþ in the presence of 150 mM Naþ. Reproduced with permission from Mayes AG, Blyth J, Millington RB, and Lowe CR (2002) Analytical Chemistry 74, 3649–3657.

25 mg/ml), dried and subsequently crosslinked and immobilized with 1.25% (v/v) glutardialdehyde. The urease-modified DMAEM holograms were exposed to various concentrations of urea up to 100 mM at pH 7.0 and the Bragg shift noted: A linear response up to a concentration of 20 mM urea resulted in a Dlmax shift of  90 nm after  17 min. The penicillin-sensitive hologram comprised photopolymerized HEMA, EDMA (5 mol%) and MAA (6 mol%) with the enzyme being directly attached to the silver (Ag0) grains via a sulfhydryl linkage. The penicillinase-functionalized MAA hologram responded to 20 mM penicillin G with a  200 nm red shift in the Bragg wavelength in  5 min (Marshall et al., 2004b). Enzyme-coupled holographic sensors are representative of a generic system whereby individual enzymes or a series of linked enzymes can be integrated with holograms to generate a family of inexpensive/disposable sensors for a wide range of biochemical metabolites. Alternatively, co-entrapment of enzymes within the holographic matrix can be used to eliminate potential interferences, monitor enzyme inhibitors or amplify weak responses to other analytes. A label-free holographic inhibition assay has been developed for drug discovery using acetylcholinesterase (EC 3.1.1.7, AChE) as an exemplar (Tan and Lowe, 2009). Holograms (8 mm dia) were fabricated in a 2x2 array using a mask for co-polymerizing HEMA, MAA (6 mol%) and EDMA (5 mol%) and the enzyme (50 U/ml) immobilized as for urease except that formaldehyde vapor was used as the crosslinking agent. This sensor was able to monitor the release of acetic acid following hydrolysis of acetylcholine by AChE; Bragg shifts of up to 145 nm were observed by the action of up to 200 U/ml AChE on 10 mM acetylcholine. No significant loss of enzyme activity occurred even after storage for 35 days in assay buffer at 4  C, allowing repeated use of a single sensor. Moreover, the sensor array was used in determining apparent inhibitor constants (Ki) for several inhibitors, including chlorpyriphos, edrophonium, parathion methyl, galanthamine, eserine, neostigmine and tacrine, and confirmed the holographic array format may be applicable to high throughput screening (Tan and Lowe, 2009). In another configuration, a holographic ELISA system was developed based on the HEMA (89 mol%), EDMA (5 mol%) and MAA (6 mol%) hologram with hemoglobin-selective antibodies immobilized on its surface (Lowe and Marshall, 2007). In a second step, antibody labelled with penicillinase was used to create a sandwich immunoassay in the presence of the analyte, hemoglobin, and detected as a blue shift in the replay wavelength by the release of protons on hydrolysis of the substrate penicillin G with the enzyme. In a further embodiment of the system, synthetic receptor systems have been exploited to create more durable sensors. For example, the use of methacrylated a-, b- and g-cyclodextrins incorporated into polymer matrices were used to quantify apolar substances such a p-nitrophenol and anthraquinone-2-carboxylate (Lowe et al., 2003a, b). The a-, b-, and g-cyclodextrin-based holographic sensors exhibited approximately 10, 14, and 19 nm of Bragg peak shift (Dlmax), respectively, in the presence of 1.0 mM anthraquinone-2 carboxylate, in pH 7.5 at 30  C. In addition, a- and g-cyclodextrins were used to quantify the concentration of p-nitrophenol, resulting in approximately 8.5 and 2.0 nm shifts, respectively, for 1 mM of the target molecule. Similarly, molecularly imprinted polymer (MIP) films have been used to create a holographic sensor for testosterone (Fuchs et al., 2013,

130

Physical Sensors: Holographic Sensors

2014). The support polymer comprised MAA as the functional monomer and EDMA as the crosslinker with a 1:1 mol% monomer/ crosslinker ratio in order to achieve adequate MIP formation. However, at these high crosslinker concentrations, the elasticity of the holographic system is severely limited and Bragg shifts of only 3.6 and 2.8 nm were achieved at 10 and 1 mM respectively. While other explanations could be found for these small Bragg shifts, the requirement for highly crosslinked polymers to ensure molecular recognition and specificity is in unbridgeable conflict with the need for flexibility in the holographic grating.

Holographic glucose sensors Early diagnosis and management of diabetes allows treatment and self-monitoring through quantification of glucose concentrations in biological fluids. It is well established that boronic acids are capable of forming reversible covalent bonds with cis-diols (Davis and Wareham, 1999; Kraiskii et al., 2010). Previous work has shown that holograms fabricated with synthetic receptors such as 4-vinylphenylboronic acid (4-VBPA) polymerized with acrylamide and N,N0 -methylenebisacrylamide as the glucosesensing moiety can detect glucose at pH 9 but not at lower pH values typical of physiological fluids (Kabilan et al., 2004a, b). However, it is also known that hydrogels constructed with 3-acrylamidophenylboronic acid (3-APB) change volume in the presence of glucose at physiological pH values (Hisamitsu et al., 1997). The ability to form reversible covalent bonds with the cis-diol groups of glucose, fructose and other carbohydrates at physiological pH is believed to be due to a decrease in the pKa of the boronic acid moiety as a result of the presence of the acrylamido group at the meta position of the phenyl ring (Norrild and Sotofte, 2002; Kabilan et al., 2005; Horgan et al., 2006). Phenylboronic acid at pH values below its pKa ( 8.8) resides in an uncharged and trigonal planar (sp2)configuration, while at higher pH values (> pKa) the trigonal form reacts with –OH groups to form a more stable negatively charged tetrahedral (sp3) state which can bind cis-diol groups more readily. Holographic grating sensors comprising polyacrylamide (pAA) crosslinked with N,N0 -methylenebisacrylamide matrixes functionalized with 3-acrylamidophenylboronic acid (3-APB) have been fabricated and their sensitivity optimized for glucose detection by varying the mol% 3-APB to achieve a maximum response at 20 mol% (Lee et al., 2004; Kabilan et al., 2005). The holographic sensor responded with a monochromatic red shift in the diffracted peak wavelength (lmax) as a function of increasing glucose concentration (2-10 mM) (Fig. 4). The binding of glucose to 3APB in the holographic matrix creates charged tetrahedral phenylboronate anions which in turn results in a Donnan pressure increase, imbibition of bulk water and swelling of the gel and, ultimately, an increase in the grating spacing and a shift to longer replay wavelengths. Glucose binding to the phenylboronate is reversible and when the hologram is placed in a glucose-free solution, the hologram contracts and the diffraction wavelength returns to its starting value. This behavior allows the sensor to track continuously dynamic changes in glucose concentration in real time (Heidari et al., 2017). Significant challenges have been encountered with deploying 3-APB as a reversible glucose receptor: its lack of specificity in physiological fluids, in that it binds to other cis-diol containing analytes, notably fructose and lactate, its dependency on ionic strength and its pH sensitivity. These features make 3-APB-based holographic sensors especially unsuitable for monitoring tear fluid and bacterial culture media, especially because lactate is a metabolic by-product that is present at high concentrations in both these fluids and their pH is highly variable. A solution has been to design new ligands that can selectively bind to glucose and obviate lactate interference (Yang et al., 2006). It is known that the tetrahedral form of the boronic acid favors binding with glucose, whereas the trigonal form preferentially binds with lactate. Thus, if it is possible to generate a tetrahedral boronate analog as the dominant species at physiological pH values, then the interference from lactate is likely to be diminished.

Fig. 4 Holographic glucose sensors based on 3-APB: (A) Formation of the tetrahedral complex and wavelength Shift as a function of glucose concentration, (B) wavelength shift as a function of glucose concentration at two scales. Reproduced with permission from Yetisen A. K. PhD Thesis. University of Cambridge.

Physical Sensors: Holographic Sensors

131

Yang et al. (2006) designed, synthesized and characterized a new boronate analog, 2-acrylamidophenylboronate (2-APB) and assessed its boron geometry by 1H, 11B and 13C NMR spectroscopy, IR and mass spectroscopy and compared its affinity towards glucose and lactate to that of 3-APB. Multiple forms of 2-APB, originating from a neighboring effect of the amido group with the boronic acid through an intramolecular B-O-coordinated interaction, were shown to exist in solution by using multinuclear NMR spectrometry. It was found that 2-APB predominantly adopts a zwitterionic tetrahedral form at physiological pH values. The complex formation of 2-APB with glucose and lactate was investigated in DMSO; 2-APB favors binding with glucose rather than lactate and generates a six-membered-ring complex. The pH titration results demonstrate that the tetrahedral forms are the dominant species present in 2-APB, whereas the trigonal and tetrahedral forms coexist in 3-APB, with the trigonal form being the dominant species at neutral pH values. A holographic sensor based on an acrylamide hydrogel, containing 2-APB as the glucose-responsive functionality incorporated with co-monomers poly(ethylene glycol) acrylate (PEG), (3-acrylamidopropyl) trimethylammonium chloride (ATMA) and [2-(acryloyloxy)ethyl]-trimethylammonium chloride (AETA) contracted on addition of glucose at physiological pH values due to the formation of a 2:1 complex between the tetrahedral 2-APB and glucose. This behavior is in contrast to the observed swelling of the 3-APB-based hologram in the presence of glucose. It is believed that recognition of glucose by 3-APB results in a 1:1 complex, increases the degree of ionization and creates a Donnan potential between the gel and bulk phases, whence subsequent swelling induces a red shift in the replay wavelength. However, the glucose-induced contraction of the 2-APB-based holographic sensor is proposed to involve the formation of a 1:2 complex promoted by the higher affinity of the boron center in its tetrahedral configuration (Ni et al., 2004; Dowlut and Hall, 2006; Zhu et al., 2006; Pan et al., 2008). Furthermore, the 2-APB-based hologram displayed a substantially reduced interference from lactate and a glucose-binding profile that was almost independent of pH over the physiological pH range. These features are desirable for development of a contact lensbased glucose sensor, where the pH variability is greater (pH 5.8–7.8) and the lactate concentration can be up to 200  higher than in blood. Furthermore, the 2-APB-based holographic sensors also displayed a fast response to glucose. The successful union of holograms and the tetrahedral 2-APB receptor for glucose detection in artificial tear fluid has also been demonstrated (Yang et al., 2006, 2008). The applicability of holographic sensors for monitoring glucose in real samples has been demonstrated. For example, the concentration of glucose was quantified in vitro with a hologram comprising 3-APB and ATMA to measure human blood plasma samples at concentrations of 3–33 mmol/L over an extended period for application in real-time continuous monitoring (Worsley et al., 2007). The results confirmed that the sensors had not only a performance comparable to that of electrochemical sensors, without lag or hysteresis, but that the measurement accuracy was not affected in the presence of common antibiotics, diabetic drugs, pain killers and endogenous substances (Worsley et al., 2007). In another study, the use of holographic sensors in urinalysis was investigated and compared with gold standard high throughput analysis and point-of-care Clinistix (Yetisen et al., 2014b). The holographic sensors showed reversible Bragg peak shifts within the range 510–1100 nm, and a performance (R2 ¼ 0.79) within 5 min well in excess of commercial colorimetric dipsticks read by automated readers (R2 ¼ 0.28) and a comparable accuracy to fully automated clinical chemistry analysers (Yetisen et al., 2014b). The holographic sensors were also tested as constituents of a minimally-invasive contact lens glucose monitoring system (Domschke et al., 2006a, b; Domschke, 2010). This sensor configuration eliminates the challenge of measuring absolute fluorescence intensities and detects relative changes in tear glucose concentration. The peak diffraction wavelength appeared to track the increasing blood glucose concentration with little or no delay. The ophthalmic glucose sensor offers a less invasive and more convenient means of continuous monitoring than other approaches and the spectrometer could be miniaturized into inconspicuous shapes, such as a make-up compact, and reflection holographic sensors could be produced cost effectively (Domschke, 2010). Nevertheless, it is clear that larger studies are needed to evaluate the possibility of long-term tear glucose monitoring with these contact lens sensors (Baca et al., 2007; Farandos et al., 2014).

The lactate challenge is an a-hydroxy acid produced as a result of anaerobic metabolism and is often a key analyte per se in the food, fermentation, clinical and sports industries. Boronic acids are known to form rapid and reversible complexes with bidentate ligands to form five- or six-membered cyclic esters. While most recent work has focused on the binding of polyols, particularly saccharides, boronic acids are also known to bind o-diphenols, o-hydroxy acids, dicarboxylic acids, and a-hydroxy acids, including L-lactate. a-hydroxy acids are believed to bind to the trigonal (sp2) form of the boronate, although reaction with the tetrahedral (sp3) form does occur, but at a significantly slower rate (Babcock and Pizer, 1980). Thus, a holographic response for L-lactate should be achieved by employing a boronate moiety as the functional ligand. The development of a holographic sensor to enable the selective and continuous real-time measurement of L-lactate has been reported using three boronic acid -based receptors (2-APB, 3-APB, 4-APB) (Sartain et al., 2006). It was found that the incorporation of 3-APB into an acrylamide hydrogel generated the largest response towards Llactate. The effects of hydrogel composition, varying L-lactate concentrations and the response of potential interfering agents to the sensor were investigated. The cross-selectivity of 3-APB holographic sensors for glucose and L-lactate presents a conundrum – how to achieve unique selectivity for these two analytes in circumstances where both naturally occur in physiological samples in various concentration ratios. This is particularly relevant for measuring L-lactate in tear fluid, where the concentration of L-lactate can be substantially higher than the glucose concentration, whereas in whole blood, the converse is true. In a later NMR and holographic study, it was shown that Llactate binds to 3- and 4-APB in aqueous solution at pH 7.4, but not to 2-APB (Sartain et al., 2008). It is evident that at pH 7.4, 3-

L-Lactate

132

Physical Sensors: Holographic Sensors

Fig. 5 Comparison of the characteristics of 2-APB and 3-APB as ligands for binding glucose and L-lactate. (A) 11B NMR as a function of pH, (B) Calibration curves for holographic 2-APB and 3-APB as a function of pH over the physiological range of tear fluid (pH 5.8–7.8), (C) Holographic responses of 2-APB to glucose and L-lactate, and glucose þ10 mM L-lactate. Adapted with permission from Yang X-P, Lee, M-C, Sartain F, Pan X, and Lowe CR (2006) Chemistry – A European Journal 12: 8491–8497.

APB predominantly resides in a trigonal configuration, whereas 2-APB more likely adopts a zwitterionic tetrahedral form, which may repel electrostatically the negatively charged lactate. Studies on the interaction of glucose and L-lactate with 5-amino-2hydroxymethylphenyl boronate by 11B NMR spectroscopy demonstrated that L-lactate bound and created a stable negatively charged tetrahedral complex, whereas addition of glucose produced little or no change in the NMR signal (Sartain et al., 2008). Thus, one would predict that when a similar boronate receptor is incorporated into a hydrogel that the resulting holographic sensor would respond selectively to L-lactate but not glucose. Similarly, 2-APB is a suitable boronate receptor for glucose and minimizes the effect of interference from L-lactate, particularly in cases where the L-lactate: glucose ratio is in favor of L-lactate, such as in tear fluid (Yang et al., 2006, 2008) (Fig. 5).

Holographic sensors for microorganisms A number of studies have now demonstrated the value of pH-sensitive holographic sensors in monitoring the growth of microorganisms (Marshall et al., 2003; Lee et al., 2004). In addition, holographic sensors for the detection of spore germination and vegetative growth in Bacillus spp. have been described (Bhatta et al., 2007). Divalent metal ion-sensitive holograms containing a methacrylated analog of nitrilotriacetic acid (NTA) as the chelating monomer were successfully used to monitor Ca2þ ions released during Bacillus subtilis spore germination in real-time (Fig. 6). The holographic response manifested as a blue-shift (16 nm) in the diffraction wavelength over the progress of germination. Similarly, pH-sensitive holograms comprising methacrylic acid (MAA) as the ionizable monomer were responsive to changes in pH associated with early vegetative metabolism following germination of Bacillus megaterium spores; a visually perceptible blue-shift (75 nm) in holographic replay wavelength was observed. Casein and starch-based holographic matrices, prepared by co-polymerization of the appropriate substrate with acrylamide, were also used to detect exo-enzymes released during later stages of Bacillus megaterium and Bacillus subtilis vegetative cell growth by observing the reduction in diffraction intensity due to progressive fringe disruption caused by enzymatic cleavage. The combined monitoring of various germination and growth events using a range of holographic sensors provides a novel, comprehensive means for the detection of viable bacterial spores. A further study created an enzyme-linked holographic sensor for the detection of spore-specific calcium dipicolinate (Ca-DPA) using small, acid-soluble spore proteins (SASPs) extracted from dormant spores as the holographic matrix (Bhatta et al., 2008). Changes in holographic diffraction intensity were used to assess the structural integrity of SASP-based matrices in response to proteolysis by recombinant germination proteases (GPRS) activated by Ca-DPA. This wok demonstrated the potential of monitoring Ca-

Physical Sensors: Holographic Sensors

133

Fig. 6 Monitoring of microbial action with holographic sensors: (i) Ethanolic fermentation of glucose with Saccharomyces cerevisiae and following ethanol production as a function of time with an ethanol-sensitive hologram and GC–MS. (ii) Homolactic fermentation of milk with Lactobacillus casei monitored with a pH-sensitive hologram and a conventional glass pH micro-electrode. (iii) Germination of Bacillus megaterium spores monitored by the release of Ca2þ ions with a Ca2þ-sensitive hologram and absorbance at 600 nm.

DPA released from germinating spores to activate GPRS proximal to a SASP-based holographic matrix and thus create a platform to detect selectively bacterial spores with intrinsic signal amplification.

Holographic sensors for physical parameters Holographic sensors that respond to physical parameters such as pressure, temperature, light intensity and magnetic field strength have been designed and reported (Lowe et al., 2006). For example, a holographic sensor that responded to pressure has been created from a co-polymer of acrylamide: methacrylamide (2:1 mol%) cross-linked with MBA (5 mol%) and shown to respond to perpendicular and horizontal compressional forces (Lowe et al., 2006; Martínez-Hurtado, 2013). However, systematic studies are required to examine a range of elastic, thermally-, magnetically- and photo-sensitive polymers and quantitatively measure the resulting effects of the physical parameters (Li et al., 2018).

New holographic procedures Measuring molecules in the natural or clinical environment, particularly gas molecules such as hydrocarbons or volatile organic compounds (VOCs), has been of growing concern in recent years. Combustible hydrocarbon gases and VOCs are released into the atmosphere from anthropogenic sources, road transport, indoor furniture and other industrial processes. Hydrophobic polymer films are known to interact with apolar molecules although there is still a requirement to incorporate these polymers into novel sensor configurations in order to exploit fully the gas-polymer molecular interactions as signal transducers, thereby converting them into readable measurements. It has been demonstrated that by using a novel diffusion method to form nanoparticles in situ, followed by laser ablation (Hajiesmaeilbaigi et al., 2005), it is possible to produce holographic sensors in materials in which volume holograms have never previously been recorded (Martinez-Hurtado et al., 2010, 2011). The approach was exemplified by recording holograms in hydrophobic films of poly(dimethylsiloxane) (PDMS) and assessing their ability to monitor hydrocarbons and other volatile compounds. The procedure entailed spreading solutions of an apolar silver salt (silver perfluoropropionate, AgPFP) in a solvent (tetrahydrofuran, THF) on the surface of PDMS films, allowing the salt and solvent to perfuse into the

134

Physical Sensors: Holographic Sensors

PDMS film, whence the salt was reduced with hydroquinone (HQ) to promote silver (Ag0) nanoparticle growth and form semitranslucent dark olive green films. Unlike previous holographic protocols, the silver nanoparticles in the polymer matrix were ablated with a Nd:YAG pulsed laser at 532 nm to form standing waves and an interference pattern, with bright fringes generated in the crests where the particles were ablated. This technique allowed a readily available hydrophobic silicone elastomer to be transformed into an effective holographic sensor for hydrocarbon gases and other volatile compounds, with the known intermolecular interactions present between the polymer and molecules being used to predict the sensor performance. The hydrophobicity of PDMS allows the sensor to operate without interference from water and other atmospheric gases and thus makes the sensor suitable for biomedical, industrial, or environmental analysis (Martinez-Hurtado et al., 2010) (Table 1). The laser ablation approach has been applied to the fabrication of other sensor holograms: For example, for the fabrication of photonic structures in Nafion membranes for the detection of ammonia in the 0.19–12.5% (v/v) range (Martinez-Hurtado and Lowe, 2014). The gratings were recorded by laser ablation of silver nanoparticles synthesized in situ by diffusion and displayed replay wavelengths which were directly proportional to the concentration of ammonia after exposure. The effective concentration range that could be measured with these membranes covers both the fatal limit of exposure and the lower flammable limit of gaseous ammonia and makes these sensors suitable for remote sensing and real-time monitoring of gases. In a further manifestation of this approach, a novel holographic platform, fabricated via laser ablation on a chitosan hydrogel with gold nanoparticles with a replay wavelength in the visible and near IR, was sensitive to glucose in the range 0–70 mM at pH 7.4 and in the presence of 154 mM salt (Vezouviou and Lowe, 2015). The sensor was unresponsive to potential interferences found in the interstitial fluid, such as fructose, vitamin C and lactate, at their respective normal concentrations, and was stable to fluctuations in temperature, pH and ionic strength. The characteristics of this sensor suggests that it may be applicable for use as an implanted device for the real time monitoring of glucose concentrations in interstitial fluid using near IR as the interrogating wavelength, possibly in wearable (Lowe, 2010), mobile healthcare (Khalili Moghaddam and Lowe, 2018), paper (Yetisen et al., 2013) and closed-loop integration of point-of-care technologies (Tan et al., 2019). Other developments in holographic sensor technologies include light-directed writing of chemically tunable narrow-band holographic sensors (Yetisen et al., 2014c), printable surface holograms via laser ablation (Vasconcellos et al., 2014), nanoparticle-free fabrication of tunable holographic sensors through pulsed laser writing (Yetisen et al., 2014b, d) and laser-generated photonic nanosensors (Yetisen et al., 2014e). Holographic optical diffusing microstructures imprinted on the surface of glucoseresponsive hydrogels through a UV-curing process have also been introduced (Elsherif et al., 2018). Volumetric changes in the hydrogel modulated the dimensions of the optical diffusing microstructures and hence the efficiency of the sensor to scatter light. This results in the beam profile and power of both the transmitted and reflected beams being changed and monitored with a power meter and smartphone app. Altering the position, angle or intensity of the diffracted beam can also be achieved by incorporating holographic optical elements into the sensor platform (Dobson, 2007). The use of a plane mirror with an offset angle of  4 creates a hologram with a narrow range of viewing angle, typically  4 , which requires fine judgement of the position of the diffracted light

Table 1

New developments in holographic sensor technologies.

New technology

References

Apolar matrices Laser ablation Near IR holographic sensors Chemically-tunable narrow-band holograms Printable surface holograms Pulsed laser writing Photonic nanosensors Optical elements Holographic optical diffusing microstructures Corner cube retroreflectors Self-assembled holographic sensors Double polymerized holograms Metal nanoparticle holographic sensors Chemical combinatorial holographic sensors Smartphone readable holograms Nanosensor arrays Modelling

Martinez-Hurtado et al. (2010, 2011) Martinez-Hurtado and Lowe (2014) Vezouviou and Lowe (2015) Yetisen et al. (2014c) Vasconcellos et al. (2014) Yetisen et al. (2014b, d) Yetisen et al. (2014e) Dobson (2007) Elsherif et al. (2018) Ahmed et al. (2017) Kwok Kwan Yuk (2012) Khalili Moghaddam et al. (2018) Versnel (2012) Oliveira et al. (2017) Khalili Moghaddam and Lowe (2017) Chan et al. (2019) Naydenova et al. (2004) Tsangarides et al. (2014) Martinez-Hurtado and Lowe (2015) Versnel et al. (2017) Fuchs et al. (2013, 2014) Cody et al. (2018) Vorzobova and Sokolov (2019)

Enhanced reflectivity holograms Molecularly imprinted receptor holograms New photopolymer materials

Physical Sensors: Holographic Sensors

135

beam. If a concave or convex mirror, corner cubes or a prism is used as the reflective object during exposure with the laser, control over the direction of the diffracted beam can be achieved to permit more facile interrogation of the hologram at wider ( 25–30 ) angles and distances (Dobson, 2007). Corner cube retroreflectors (CCRs) have applications in sensors, image processing, free space communication and wireless networks. An  10 mm-thick holographic corner cube retroreflector (HCCR) array that acted as a colorselective wavelength filter and diffracted light at broad angles was also created (Ahmed et al., 2017). Angle-resolved spectral measurements showed that the Bragg peak of the diffracted light from the HCCR array could be tuned from 460 to 545 nm by varying the incident angle. The utility of the HCCR was demonstrated as optical temperature and relative humidity sensors that produced a visible colorimetric response for rapid diagnostics (Ahmed et al., 2017). The concept of a simple and direct formation of reflection gratings by the self-assembly of symmetric diblock copolymers has been demonstrated (Kwok Kwan Yuk, 2012). A diblock copolymer polystyrene-block-poly(2- and 4-)vinylpyridine (PS-b-PVP) of three molecular weights: PS-b-P2VP 198 kg/mol-b-192 kg/mol, 213 kg/mol-b-206 kg/mol and PS-b-P4VP 195 kg/mol-b-204 kg/ mol was employed in the study. The addition of 3-pentadecylphenol (PDP) into the diblock copolymers conferred noticeable improvement in the grating quality and the optical features of the gratings were found to be in good agreement with the expected lamellar dimensions of the diblock copolymer. Chemical functionalities such as protonatable groups were introduced to the diblock copolymer gratings PS-b-P2VP (213 kg/mol-b-206 kg/mol) to render additional sensing-capability to the reflection grating (Kwok Kwan Yuk, 2012). A more recent report resolves an issue with the use of silver nanoparticles in holographic sensors in contact lenses, i.e. that silver nanoparticles require a complex regulatory approval process by the FDA, are aesthetically unappealing in that they appear as a dark colored visible spot in the contact lenses and would require extensive modification of existing lens manufacturing lines (Domschke et al., 2006a, b). The new approach creates an alternative metal-free, transparent glucose-sensitive holographic grating using a double polymerization protocol (Khalili Moghaddam et al., 2018). Modulation of the refractive index is achieved by polymerizing a second more highly cross-linked polymer (P2) within the first baseline polymer (P1). The polymerization of P2 is a function of the standing wave, where, in the light regions, the exposure strongly promotes polymerization and, consequently, a sinusoidal concentration profile of P2 is formed that modulates the permittivity of the polymer material and generates a grating structure. For holographic sensor applications, the P1 phase is functionalized with 3-APB to develop a glucose-responsive “smart” hydrogel. However, the response of the sensors was complicated by the fact that recording the holographic gratings using a highly crosslinked polymer (P2) within the lattice of a low cross-linked polymer (P1) created two functionalized phases which may have displayed different binding kinetics. The wavelength response of the 3-APB-based double polymerized sensor showed a biphasic time-dependent expansion-contraction profile in contrast to the expansion only behavior of the conventional 3-APB-based metal nanoparticle sensors. Nevertheless, the double polymerization approach is robust and generates a transparent final product which can be configured for use in daily-wear contact lenses. Samples of the double-polymerized holographic sensors were developed and preliminary studies on the response to glucose performed.

Chemical and combinatorial holography Altering the chemistry of the holographically diffracting nanoparticles can also be used to change the selectivity of the holographic sensor. For example, not only can silver (Ag0) nanoparticles be exchanged for gold (Au0) particles (Versnel, 2012; Vezouviou and Lowe, 2015), but bright Denisyuk holograms can be fabricated by ablation of poly(2-hydroxyethyl methacrylate films doped with iron, nickel and cobalt nanoparticles and exploited to monitor gaseous nitric oxide (NO) (Versnel, 2012). The prospect of combining combinatorial chemistry with holography potentially broadens the scope of analytes that can be monitored with holographic sensors and has been used to design and develop an inexpensive sensor for detecting illicit drugs (Oliveira et al., 2017). The increasing consumption of illicit drugs is a global social and medical challenge and eradicating the supply and use of these drugs necessitates the development and deployment of devices that provide a fast, selective and simple response in seized powder samples. A novel strategy for the development of a holographic sensor with a rationally designed receptor for the detection of cocaine has been demonstrated. The receptor was designed by identifying known protein–cocaine interactions, particularly those of human carboxylesterase-1 and catalytic monoclonal antibody GNL7A1, and constructing a four-component Ugi library of putative cocaine-binding ligands using a hybrid hydrogel of poly (2-hydroxyethyl methacrylate–co-ethylene dimethacrylate-co-methacrolein) as the carbonyl source for the Ugi reaction. A Bragg diffraction grating comprising gold nanoparticles was fabricated in the polymer gel and the replay wavelength of the combinatorial library of Ugi-modified holographic sensors measured via a spectrophotometer in response to cocaine as a function of pH and ionic strength. A smartphone-based instrument and cloudbased algorithm was employed (Khalili Moghaddam and Lowe, 2017), optimised with training and test samples of cocaine and used to assess the presence of cocaine in seized field samples. These work confirmed the potential application of this novel analytical approach in field testing for illicit drugs (Fig. 7).

Miniaturization of holographic sensors There is a strong demand in microbiological, cellular and pharmaceutical research for tools which permit real-time analysis of analytes in small volume samples. Lowe et al. (2004) hypothesized that 3 mm “spot” holographic sensors could be fabricated using

136

Physical Sensors: Holographic Sensors

Fig. 7 Combinatorial holography based on the 4-component Ugi reaction for the quantification of Illicit drugs via a smartphone system. Reproduced with permission from Oliveira et al. (2018) Sensors and Actuators B: Chemical 270: 216–222.

a contact printing process involving a hydrophobic FEP mask to inhibit polymerization at the masked regions where it was oxygen rich (Lowe et al., 2004). Low impact holographic sensors have been incorporated into inexpensively fabricated PDMS microfluidic chambers to allow unimpeded growth of microorganisms in small volumes without interruption due to interrogation (Bell et al., 2010), using as an exemplar for proof-of-concept, the pH change caused by the real-time homolactic fermentation of Lactobacillus casei. A tape mask was used to a tape mask to fabricate a spot holographic sensor of area about 1 cm2 (10 mL). A key challenge for the miniaturization of the pH holographic sensor is that the gel volume relative to the bioreactor volume and maximizing the holographic sensor signal without adding to the buffering capacity of the system. In a further study, small volume (3 nL) nonconsumptive holographic sensors were integrated into a glass-PDMS microfluidic chip to monitor the blue shift in the diffraction replay wavelength of a pH-sensitive holographic sensor during the microbial growth of Lactobacillus casei Shirota (Chan et al., 2019) (Fig. 8). Within the optimal growth pH range of L. casei, the accuracy of the miniaturized pH sensors was comparable to that of a gold standard pH meter. Conceivably, this approach could be extrapolated to an array of miniaturized holographic sensors sensitive to different analytes, and thereby paving the way for reliable, real-time, non-invasive monitoring of microorganisms in a nanobioreactor.

Modelling and quantitation of holographic sensors A number of modelling studies have been undertaken in order to understand, and in some improve, the performance of the holographic sensors. In one study, a pH-sensitive silver nanoparticle holographic sensor was fabricated through photochemical patterning and its behavior predicted through computational simulations through a finite element modelling technique in order to analyze its optical properties on varying the pattern and characteristics of the nanoparticle arrays within the responsive hydrogel matrix. The simulations and the wavelength tunability of the sensor were in good agreement with the experimental results. Various factors, including nanoparticle size and distribution, within the hydrogel-based responsive matrices directly affect the performance of the sensors (Tsangarides et al., 2014). While conventional volume holographic gratings (VHG) fabricated in photosensitive emulsions such as gelatine containing silver salts are easily visualized in ambient lighting, for holographic sensors which require a more defined and chemically-

Physical Sensors: Holographic Sensors

137

Fig. 8 Microarray of holographic pH sensors (A) 1 mm dia; acrylic acid functional monomer), (B) scanning image of an individual microhologram and (c) the replay wavelength (lmax) of all 9 microholograms as a function of pH in the range 3.5–8.0.

functionalized polymer matrix, laser ablation has been introduced. However, the replay signal from these holograms can be challenging to identify in ambient lighting, particularly when traditional photochemical bleaching solutions are used to reduce light scattering and modulate the refractive index within the laser ablated volume holographic gratings, the holographic peak intensity is often reduced. This is postulated to occur because both light and dark fringes contain a proportion of metal particles, which upon solubilization are converted immediately to silver iodide, yielding no net refractive index modulation. A recent study advances the hypothesis that the reduced intensity of holographic replay signals is linked to a gradient of different sized metal particles within the emulsion, which reduces the holographic signal and may explain why traditional bleaching processes result in a reduction in intensity (Versnel et al., 2017). A novel experimental protocol was provided, along with simulations based on an effective medium periodic 1D stack, that offered a solution to increase peak signal intensity of holographic sensors by greater than 200%. Nitric acid was used to etch the silver nanoparticles within the polymer matrix and is believed to remove the smaller particles to generate more defined metal fringes containing a soluble metal salt. Once the grating efficiency has been increased, this salt can be converted to a silver halide, to modulate the refractive index and increase the intensity of the holographic signal. This new protocol has been tested in a range of polymer chemistries and it was concluded that those containing functional groups that stabilize the metal nanoparticles within the matrix yield more intense holographic signals as the integrity of the fringe is more protected with increasing metal solubility (Versnel et al., 2017). The photonic phenomena associated with holographic sensors are relatively well understood and have been described theoretically using various modelling methodologies (Naydenova et al., 2004; Martinez-Hurtado et al., 2010; Tsangarides et al., 2014). However, a theoretical model that considers both photonic effects and complex diffusion of the analytes into the gel phase has only been proposed recently for these systems (Martinez-Hurtado and Lowe, 2015). Molecular transport into the polymer matrix is governed by diffusion, which appears to become anomalous in the presence of nano-particles, with interactions between the diffusant, the polymer and the nanoparticles all affecting this process (Liu and de Kee (2005a). Thus, diffusion models coupled with rheology have been proposed for modelling these complex processes in polymeric membranes (Liu and de Kee, 2005b). Further extensions of these models account for swelling, consider microscopic (kinetic) and mesoscopic (thermodynamic) levels, energies of mixing, molecular polymer structure conformations, interfacial interactions (Puri et al., 2008) and molecular dynamics simulations for transport processes (Xiao et al., 2012). Martinez-Hurtado and Lowe (2015) discuss diffusion, swelling, mixing, and changes in optical diffraction as elements of an integrated theoretical model. A significant advantage of holographic sensors is that they provide a real-time color or alphanumeric image response to the analyte of interest which is readily readable and interpretable by the human eye. However, a visual assessment of a holographic sensor by an inexperienced individual is often only able to record a semi-quantitative interpretation of analyte concentrations in categories such as positive, negative, high or low. Improving the sensitivity and quantification of the color characteristics requires the implementation of more sophisticated measurement instruments. The diffraction wavelength of the holographic sensors has been most commonly analyzed quantitatively using an Avantes AVS-MC2000-2 reflectance spectrometer and AvaSoft 5 processing software available from Knight Scientific (UK) (Blyth et al., 1996; Kabilan et al., 2004a, b; Lowe, 1999; Lowe, 2007; Naydenova et al.,

138

Physical Sensors: Holographic Sensors

Fig. 9 Use of the smartphone to capture, encrypt, identify and quantitate holographic sensor responses. Reproduced under the Creative Commons Attribution License from Khalili Moghaddam G and Lowe CR (2017) PLoS One 12(11): e0187467.

2008; Khalili Moghaddam et al., 2018). However, while colorimetric tests have been monitored with the camera system found in smartphones (Yetisen et al., 2014f) with good reproducibility, only recently have smartphones been used to quantify holographic sensors (Khalili Moghaddam and Lowe, 2017) (Fig. 9). The value of this approach was exemplified by analyzing the replay color of captures images of a pH-sensitive hologram in real-time. It was achieved by personalized encryption and a wavelet-based image compression method to secure image transfer to the cloud, whence the decrypted and decompressed image was processed via four principal operations: Recognition of the hologram in the captured image within a complex background using a templatebased approach, conversion of device-dependent RGB values to device-independent CIEXYZ values by implementing a polynomial model of the camera and computation of the CIEL*a*b* values, using the color coordinates of the image to segment the image and select the appropriate color descriptors, ultimately to locate the holographic region of interest (ROI) and, finally, applying a machine learning algorithm to correlate the color coordinates of the ROI to the analyte concentration. A key step was identifying the sensor output in a captured image based on a color cue, and while this was achieved, in the real world this might be obscured by complex backgrounds. Thus, sensor recognition is a complex scene requires incorporating other features such as object recognition algorithms. A pattern-based technique using a QR code incorporated in the design of the sensor was explored in this work (Khalili Moghaddam and Lowe, 2017). The information in the QR code could be personalized for the patient to provide a unique identification number for the electronic health record. Thus, integrating holographic sensors, QR codes and color image processing would create a cost-effective platform for monitoring analytes in real-time by minimally-trained personnel.

Conclusions Holographic sensors are an exciting and versatile new addition to the analysts armamentarium but are still in an early stage of development and deployment. Compared to other optical sensor technologies, holographic sensors offer an inexpensive and up-scalable platform, mass producibility, color, two- and three-dimensional image and alpha-numeric capability, co-incorporation with other optical elements such as lenses and mirrors, miniaturizability, arrayability, qualitative and quantitative assessment both visually and using smartphones, and the ability for configuration into strips, free-standing flakes, QR codes, contact lenses and tattoos

Physical Sensors: Holographic Sensors

139

(Khalili Moghaddam and Lowe, 2017, 2018; Khalili Moghaddam et al., 2018). At present, they do suffer from relatively low selectivity for the target analyte, although the introduction of novel molecular imprinting (Fuchs et al., 2013, 2014), combinatorial chemistries (Oliveira et al., 2017), enhanced reflectivity (Versnel et al., 2017), new photopolymer materials (Cody et al., 2018; Vorzobova and Sokolov, 2019), improved modelling (Martinez-Hurtado and Lowe, 2015) and micro array technologies (Chan et al., 2019) into the armory is bound to improve their selectivity, sensitivity and general acceptability. However, introduction of these new chemical and physical approaches do little to further our understanding of the fundamental time-dependent dynamic processes involved in the holographic transduction process in hydrogel or hydrophobic support matrices. A fuller understanding of the diffusion of analytes into complex three-dimensional matrices and their reversible binding kinetics that lead to swelling or contraction and hence changes in the diffraction patterns of interrogating light beams needs to be achieved if these devices are to be used in continuous real-time environments. This is particularly evident for monitoring high molecular mass analytes such as proteins where diffusion and non-specific interactions with the polymer matrix and nanoparticle diffractors plays a significant role. Nevertheless, the holographic sensor platform is one of the most versatile of any sensor platform and can be configured to create visual or instrumentally monitored responses in discreet or real-time with changes in color, brightness, image/alphanumeric messages, position and at macro, micro and nano scales.

References Ahmed, R., Yetisen, A.K., Hyun, Y.S., Butt, H., 2017. Light: Science & Applications 6, e16214. https://doi.org/10.1038/lsa.2016.214. Babcock, L., Pizer, R., 1980. Inorganic Chemistry 19, 56–61. Baca, J.T., Finegold, D.N., Asher, S.A., 2007. The Ocular Surface 5, 280–293. Bell, L.L., Seshia, A.A., Davidson, C.A.B., Lowe, C.R., 2010. Procedia Engineering 5, 1352–1355. Benton SA (1972) Method for Making Reduced Bandwidth Holograms. U.S. Patent 3,633,989 A, January 11, 1972. Benton, S.A., Bove, V.M., 2007. Holographic Imaging. John Wiley & Sons, Hoboken, p. 173, 2007. Benton SA and Walker JL (1991) Holographic Colour Control Systems. U.S. Patent 4,986,619 A, January 22, 1991. Bhatta, D., Christie, G., Madrigal-Gonzalez, B., Blyth, J., Lowe, C.R., 2007. Biosensors & Bioelectronics 23, 520–527. Bhatta, D., Christie, G., Lowe, C.R., 2008. Sensors and Actuators B: Chemical 134, 356–359. Blyth, J., Mayes, A.G., Frears, E.R., Millington, R.B., Lowe, C.R., 1996. Analytical Chemistry 68, 1089–1094. Blyth J, Lowe CR, Mayes AG, and Millington RB (1999a) Holographic Sensors and their Production. WO Patent Application 1999063408 A1, December 9, 1999. Blyth, J., Millington, R.B., Mayes, A.G., Lowe, C.R., 1999b. The Imaging Science Journal 47, 87–91. Born, M., Wolf, E., 1999. Principles of Optics: Electromagnetic Theory of Propagation, Interference and Diffraction of Light, 7th edn. Cambridge University Press, Cambridge, p. 296. 1999. Bryngdahl, O., 1972. Applied Optics 11, 195–199. Chan, L.C.Z., Khalili Moghaddam, G., Wang, Z., Lowe, C.R., 2019. ACS Sensors 4, 456–463. Cody, D., Gul, S., Mikulchyk, T., Irfan, M., Kharchenko, A., Goldyn, K., Martin, S., Mintova, S., Cassidy, J., Naydenova, I., 2018. Applied Optics 57, 173–183. Collier, R.J., Burckhardt, C.B., Lin, L.H., 1971. Optical Holography. Academic Press, New York, p. 152, 1971. Davis, A.P., Wareham, R.S., 1999. Angewandte Chemie, International Edition 38, 2976–2996. Denisyuk, Y.N., 1962. Doklady Akademii Nauk SSSR 144, 1275–1279. Dobson, C.A., 2007. Design, Fabrication and Characterization of Holographic Optical Elements. Ph.D. The University of Cambridge, Cambridge, UK, 2007. Domschke, A., 2010. Chimia 64, 43–44. Domschke, A., March, W.F., Kabilan, S., Lowe, C.R., 2006a. Diabetes Technology & Therapeutics 8, 89–93. Domschke, A., Kabilan, S., Anand, R., Caines, M., Fetter, D., Griffith, P., James, K., Karangu, N., Smith, D., Vargas, M., Zeng, J., Hussain, A., Yang, X.-P., Blyth, M.A., Herbrechtsmeier, P., Lowe, C.R., 2006b. Proceedings of IEEE 3, 1320–1323. Dowlut, M., Hall, D.G., 2006. Journal of the American Chemical Society 128, 4226–4227. Elsherif, M., Hassan, M.U., Yetisen, A.K., Butt, H., 2018. ACS Nano 12, 2283–2291. Farandos, N.M., Yetisen, A.K., Monteiro, M.J., Lowe, C.R., Yun, S.H., 2014. Advanced Healthcare Materials. https://doi.org/10.1002/adhm.201400504. Flory, P.J., 1953. Principles of Polymer Chemistry. Cornell University Press, Ithaca, New York, 1953. Fuchs, Y., Soppera, O., Mayes, A.G., Haupt, K., 2013. Advanced Materials 25, 566–570. Fuchs, Y., Kunath, S., Soppera, O., Haupt, K., Mayes, A.G., 2014. Advanced Functional Materials 24, 688–694. Gabor, D., 1948. Nature 161, 777–778. Hajiesmaeilbaigi, F., Mohammadalipour, A., Sabbaghzadeh, J., Hoseinkhani, S., Fallah, H., 2005. Laser Physics Letters 3, 252–256. Hariharan, P., 2010. Basics of Interferometry. Academic Press, San Diego, CA, p. 111. Heidari, M., Dabirmanesh, B., Mahdavi, H., Khajeh, K., 2017. Iranian Biomedical Journal 21, 347–348. Hisamitsu, I., Kataoka, K., Okano, T., Sakuri, Y., 1997. Pharmaceutical Research 14, 289–293. Horgan, A., Marshall, A., Kew, S., Dean, K., Creasey, C., Kabilan, S., 2006. Biosensors & Bioelectronics 221, 1838–1845. Kabilan, S., Marshall, A.J., Blyth, J., Hussain, A., Yang, X.-P., Lee, M.-C., Lowe, C.R., 2004a. Proceedings of IEEE 2, 1003–1006. Kabilan, S., Blyth, J., Lee, M.-C., Marshall, A.J., Hussain, A., Yang, X.-P., Lowe, C.R., 2004b. Journal of Molecular Recognition 17, 162–166. Kabilan, S., Marshall, A.J., Sartain, F.K., Lee, M.-C., Hussain, A., Yang, X.-P., Blyth, J., Karangu, N., James, K., Zeng, J., Smith, D., Domschke, A., Lowe, C.R., 2005. Biosensors & Bioelectronics 20, 1602–1610. Khalili Moghaddam, G., Lowe, C.R., 2017. PLoS One 12 (11) e0187467. Khalili Moghaddam, G., Lowe, C.R., 2018. Springer Briefs in Applied Sciences and Technology. Computational Intelligence. Khalili Moghaddam, G., Margerison, H., Suzuki, J., Blyth, J., Lowe, C.R., 2018. Sensors and Actuators B: Chemical 267, 1–4. Kraiskii, A.V., Postnikov, V.A., Sultanov, T.T., Khamidulin, A.V., 2010. Quantum Electronics 40, 178–182. Kwok Kwan Yuk, S., 2012. Self-Assembled Reflection Gratings. Ph.D. thesis. University of Cambridge, UK, 2012. Lee, M.-C., Kabilan, S., Hussain, A., Yang, X.-P., Blyth, J., Lowe, C.R., 2004. Analytical Chemistry 76, 5748–5755. Leith, E.N., Upatnieks, J., 1962. Journal of the Optical Society of America 52, 1123–1130. Leubner, I., Jagannathan, R., Wey, J., 1980. Photographic Science and Engineering 24, 268–272. Li N, Lowe CR, and Stevenson AC (2018) arXiv:1801.10569v1 [physics.app-ph] Lippmann, G., 1891. CRAS (Paris) 112, 274–275.

140

Physical Sensors: Holographic Sensors

Lippmann, G., 1894. Journal of Physiology, Paris 3, 97–107. Liu, Q., de Kee, D., 2005a. Journal of Non-Newtonian Fluid Mechanics 131, 32–43. Liu, Q., de Kee, D., 2005b. Rheologica Acta 44, 287–294. Lowe, C.R., 1999. Current Opinion in Chemical Biology 3, 106–111. Lowe, C.R., 2007. Holographic sensors. In: Marks, R.S., Cullen, D.C., Karube, I., Lowe, C.R., Weetall, H. (Eds.), Handbook of Biosensors and Biochips. Wiley Interscience, Sussex, UK, pp. 587–596. Lowe, C.R., 2010. Point-of-care testing. In: Price, C.P., St. John, A., Kricka, L. (Eds.), Needs, Opportunity and Innovation, 3rd edn. AACC Press. Lowe, C.R., Larbey, C., 2008. Holography Gets Smart. Physics World. February 2008. Lowe CR and Marshall AJ (2007) Use of Holographic Sensor. WO Patent Application 2007023282 A1, March 1, 2007. Lowe CR, Millington RB, Blyth J, and Mayes AG (1995) Hologram Used as a Sensor. WO Patent Application 1995026499 A1. Lowe CR, Blyth J, Marshall AJ, James A, Kabilan S, Lee MC, Madrigal-Gonzalez B, Yang X-P, and Davidson CAB (2003a) Holographic Biosensors SPIE OE magazine March 2003, 2023. Lowe CR, Davidson CAB, Blyth J, Kabilan S, Marshall AJ, Madrigal Gonzalez B, and James AP (2003b) Method of Detecting an Analyte in a Fluid. WO Patent Application 2003087899 A1, October 23, 2003. Lowe CR, Davidson CAB, Blyth J, Marshall AJ, and James AP (2004) Holographic Sensors and their Production. International Patent WO 2004081546A1, 2004 Lowe, C.R., Blyth, J. & James, A.P. (2006) Interrogation of a Sensor. WO Patent Application 2006008531 A1, January 26, 2006. Madrigal-González, B., Christie, G., Davidson, C.A.B., Blyth, J., Lowe, C.R., 2005. Analytica Chimica Acta 528, 219–228. Marshall, A.J., Blyth, J., Davidson, C.A.B., Lowe, C.R., 2003. Analytical Chemistry 75, 4423–4431. Marshall, A.J., Young, D.S., Kabilan, S., Hussain, A., Blyth, J., Lowe, C.R., 2004a. Analytica Chimica Acta 527, 13–20. Marshall, A.J., Young, D.S., Blyth, J., Kabilan, S., Lowe, C.R., 2004b. Analytical Chemistry 76, 1518–1523. Marshall, A.J., Kabilan, S., Blyth, J., Lowe, C.R., 2006. Journal of Physics. Condensed Matter 18, S619–S626. Martínez-Hurtado, J.L., 2013. Gas-Sensitive Holographic Sensors (Doctoral thesis). https://doi.org/10.17863/CAM.16072. Martinez-Hurtado, J.L., Lowe, C.R., 2014. ACS Applied Materials & Interfaces 6, 8903–8908. Martinez-Hurtado, J.L., Lowe, C.R., 2015. Journal of Membrane Science 495, 14–19. Martinez-Hurtado, J.L., Davidson, C.A.B., Blyth, J., Lowe, C.R., 2010. Langmuir 26, 15694–15699. Martinez-Hurtado, J.I., Davidson, C.A.B., Lowe, C.R., 2011. Proceedings of SPIE 8024, 80240Y. https://doi.org/10.1117/12.887295. Mayes, A.G., Blyth, J., Millington, R.B., Lowe, C.R., 1998. Journal of Molecular Recognition 11, 168–174. Mayes, A.G., Blyth, J., Kyröläinen-Reay, M., Millington, R.B., Lowe, C.R., 1999. Analytical Chemistry 71, 3390–3396. Mayes, A.G., Blyth, J., Millington, R.B., Lowe, C.R., 2002. Analytical Chemistry 74, 3649–3657. Millington, R.B., Mayes, A.G., Blyth, J., Lowe, C.R., 1995. Analytical Chemistry 67, 4229–4233. Millington, R.B., Mayes, A.G., Blyth, J., Lowe, C.R., 1996. Sensors and Actuators B: Chemical 33, 55–59. Naydenova, I., Jallapuram, R., Howard, R., Martin, S., Toal, V., 2004. Applied Optics 43, 2900–2905. Naydenova, I., Jallapuram, R., Toal, V., Martin, S., 2008. Applied Physics Letters 92, 031109. Ni, W., Kaur, G., Springsteen, G., Wang, B., Franzen, S., 2004. Bioorganic Chemistry 32, 571–581. Norrild, J.C., Sotofte, I., 2002. Journal of the Chemical Society, Perkin Transactions 1 22, 303–311. Oliveira, N.C.L., El Khoury, G., Versnel, J.M., Khalili Moghaddam, G., Sampaio, L.L., Lima Filho, J.L., Lowe, C.R., 2017. Sensors and Actuators B: Chemical 270, 216–222. Pan, X., Yang, X.-P., Lowe, C.R., 2008. Journal of Molecular Recognition 21, 205–209. Puri, P., Liu, Q., de Kee, D., 2008. Journal of Applied Polymer Science 108, 47–51. Sartain, F.K., Yang, X.-P., Lowe, C.R., 2006. Analytical Chemistry 78, 5664–5670. Sartain, F.K., Yang, X.-P., Lowe, C.R., 2008. Chemistry – A European Journal 14, 4060–4067. Saxby, G., 2004. Practical Holography, 3rd edn. Institute of Physics Publishing, London, p. 46. 2004. Seidel, J.M., Malmonge, S.M., 2000. Materials Research 3, 79–83. Spooncer, R.C., Al-Ramadhan, F.A., Jones, B.E., 1992. International Journal of Optoelectronics 7, 449–452. Tan, E.V., Lowe, C.R., 2009. Analytical Chemistry 81, 7579–7589. Tan, E.K.W., Au, Y.Z., Khalili Moghaddam, G., Occhipinti, L.G., Lowe, C.R., 2019. Trends in Biotechnology 37, 775–788. Tsangarides, C.P., Yetisen, A.K., Vasconcellos, F., Montelongo, Y., Qasim, M.M., Lowe, C.R., Wilkinson, T.D., Butt, H., 2014. RSC Advances 4, 10454–10461. Vasconcellos, F.C., Yetisen, A.K., Montelongo, Y., Butt, H., Grigore, A., Davidson, C.A.B., Blyth, J., Monteiro, M., Wilkinson, T.D., Lowe, C.R., 2014. ACS Photonics 1, 489–495. Versnel, J.M., 2012. A Holographic Nitric Oxide Sensor for Exhaled Breath Analysis in Asthma. Ph.D. The University of Cambridge, Cambridge, UK, 2012. Versnel, J.M., Williams, C., Davidson, C.A.B., Wilkinson, T.D., Lowe, C.R., 2017. Optical Materials 73, 400–407. Vezouviou, E., Lowe, C.R., 2015. Biosensors & Bioelectronics 68, 371–381. Vilkomerson, D.H.R., Bostwick, D., 1966. Applied Optics 6, 1270–1272. Vorzobova, N., Sokolov, P., 2019. Polymers 11, 2020–2034. Wen, S., Yin, Y., Stevenson, W.T.K., 1991. Journal of Applied Polymer Science 42, 1399–1406. Worsley, G.J., Tourniaire, G.A., Medlock, K.E., Sartain, F.K., Harmer, H.E., Thatcher, M., Horgan, A.M., Pritchard, J., 2007. Clinical Chemistry 53, 1820–1826. Xiao, Y., Yuan, J., Sundén, J.B., 2012. Journal of the Electrochemical Society 159, B251–B258. Yang, X.-P., Lee, M.-C., Sartain, F., Pan, X., Lowe, C.R., 2006. Chemistry – A European Journal 12, 8491–8497. Yang, X.-P., Pan, X., Blyth, J., Lowe, C.R., 2008. Biosensors & Bioelectronics 23, 899–905. Yetisen, A.K., Akram, M.S., Lowe, C.R., 2013. Lab on a Chip 13, 2210–2251. Yetisen, A.K., Naydenova, I., da Cruz Vasconcellos, F., Blyth, J., Lowe, C.R., 2014a. Chemical Reviews 114, 10654–10696. Yetisen, A.K., Montelongo, Y., da Cruz Vasconcellos, F., Martinez-Hurtado, J.L., Neupane, S., Butt, H., Qasim, M.M., Blyth, J., Burling, K., Carmody, J.B., Evans, M., Wilkinson, T.D., Kubota, L.T., Monteiro, M.J., Lowe, C.R., 2014b. Nano Letters 14, 3587–3593. Yetisen, A.K., Butt, H., Vasconcellos, F., Montelongo, Y., Davidson, C.A.B., Blyth, J., Chan, L., Carmody, J.B., Vignolini, S., Steiner, U., Baumberg, J.J., Wilkinson, T.D., Lowe, C.R., 2014c. Advanced Optical Materials 2, 250–254. Yetisen, A.K., Qasim, M.M., Nosheen, S., Wilkinson, T.D., Lowe, C.R., 2014d. Journal of Materials Chemistry C 2, 3569–3576. Yetisen, A.K., Montelongo, Y., Farandos, N.M., Naydenova, I., Lowe, C.R., Yun, S.H., 2014e. Applied Physics Letters 105, 1106. Yetisen, A.K., Martinez-Hurtado, J.L., Garcia-Melendrez, A., Vasconcellos, F., Lowe, C.R., 2014f. Sensors and Actuators B: Chemical 196, 156–160. Zhu, L., Shabbir, S.H., Gray, M., Lynch, V.M., Sorey, S., Anslyn, E.V., 2006. Journal of the American Chemical Society 128, 1222–1232.

Physical Sensors: Radiation Sensors Douglas S. McGregor and J. Kenneth Shultis, Kansas State University, Mechanical and Nuclear Engineering, Manhattan, KS, United States © 2023 Elsevier Ltd. All rights reserved.

Types of ionizing radiation Charged particles Alpha particles Beta particles, positrons, electrons, Auger electrons, conversion electrons Fission fragments and products Electromagnetic radiation X-rays Gamma rays Annihilation photons Neutral particles Neutrons Neutrinos and anti-neutrinos Gas-filled detectors General operation Ion chambers Proportional counters Geiger-Müller counters Scintillation detectors Inorganic scintillators Organic scintillators Light collection Photomultiplier tubes MicroChannel plates Photodiodes Silicon photomultipliers Semiconductor detectors Ge detectors Si detectors Alternative semiconductor detectors Personnel dosimeters Photographic film Pocket ion chambers TLDs and OSLs Alternative detectors Cloud chambers, bubble chambers, and superheated drop detectors Cryogenic detectors Radiation counting Types of measurement uncertainties Uncertainty assignment based upon counting statistics Dead time Summary References

142 142 142 142 142 143 143 143 143 143 143 144 144 144 145 145 146 147 147 149 150 150 151 151 151 152 153 154 155 155 155 155 156 156 156 157 157 157 157 159 159 159

Abstract Ionizing radiation is detected primarily by sensing the presence of liberated free charges in a control medium. The medium may be a gas, liquid, or solid with electrodes positioned about the control volume. Summarized in this article are brief descriptions of the more common ionizing radiations encountered in the environment and the nuclear industry, along with the more common methods used to detect these radiations. Radiation detectors described include gas-filled, scintillation, and semiconductor devices. A brief section is included on radiation counting statistics.

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00157-6

141

142

Physical Sensors: Radiation Sensors

Types of ionizing radiation Radiation is the emission of energy from a source, be it natural or stimulated. Although there are many types of radiation, this section is concerned mainly with the detection of ionizing radiation, consisting of energetic emissions of particles capable of producing (directly or indirectly) ionization in a substance. Typically, they are classified as charged-particle, electromagnetic, or neutral radiations.

Charged particles Alpha particles Alpha particles were discovered in 1899 by Sir Ernest Rutherford who observed emissions from a uranium sample with an electroscope (Rutherford, 1899). He deduced that at least two types of particles were emitted from the samples, one easily stopped with thin aluminum foils which he called alpha radiation and the other significantly more penetrating which he named beta radiation. Alpha particles, denoted by the Greek letter ‘a’, are doubly positive-charged particles composed of two neutrons and two protons. Although identical to helium nuclei, they are identified as particles emitted from the nucleus of an atom. By contrast, a helium atom stripped of its electrons in an accelerator is best described as a doubly-ionized helium nucleus and not an a particle. Alpha particles, with a mass of about 4.01506 u (6.644657  10 27 kg), are highly ionizing. Alpha particles are emitted from nuclei with distinctive discrete energies, which enable spectroscopic identification of a-particle emitting radioisotopes. Ranges in matter vary, but are generally short because of the high stopping power ( dE/dx) in materials, varying from a few centimeters in gases at 1 atm pressure to a few tens of microns in solids and liquids. Detection methods usually rely on sensing the free ionization produced in the detecting medium.

Beta particles, positrons, electrons, Auger electrons, conversion electrons Beta particles, also discovered in 1899 by Sir Ernest Rutherford along with a particles, were found to be generally more penetrating than a radiation. By using the method developed by J.J. Thomson to measure the mass-to-charge (m/q) ratio for electrons (Thomson, 1897), beta particles, denoted by the Greek letter ‘b,’ or more precisely ‘b’, were shown by Antoine Henri Becquerel to be fast moving electrons emitted from the nucleus of an atom (Becquerel, 1900). With a rest mass of 9.10938  10 31 kg, a b particle is an electron emitted in radioactive decay of a nucleus and is accompanied by simultaneous emission of an antineutrino emission. Consequently, b particles are emitted with a continuum of energies between zero and the Q-value of the decay reaction. A conversion electron is produced when an excited nucleus transfers its excitation energy to an orbital electron, usually in the K shell, thereby, ejecting it from the atom. This process is referred to as internal conversion. An Auger electron is similar, but is caused by an orbital electron dropping into a lower vacant electron state. The energy difference between the two states is either emitted as an X-ray (fluorescence) or transferred to an outer orbital electron, thereby, ejecting it from the atom as an Auger electron. Conversion electrons and Auger electrons are emitted with discrete energies. A positron is an antielectron, with the same rest mass and spin but opposite charge of an electron, and is emitted in a bþ radioactive decay, together with a neutrino, when a nuclear proton is transformed to a neutron through the weak force. They can also be produced in so-called pair-production photon interactions within the intense electric field near the nucleus of an atom if the photons have an energy exceeding 2mec2 ¼ 1.022 MeV (Evans, 1955; Kaplan, 1962; Shultis and Faw, 2000). They can also be produced far less frequently in the electric field of an orbital electron in a reaction called triplet production if the incident photon has an energy above 4mec2 (Evans, 1955). Positrons usually quickly annihilate with ambient electrons after they lose their kinetic energy. The annihilation produces two opposite moving photons each with an energy of 511 keV. Although much less ionizing than a particles, i.e., they have a smaller collisional stopping power and therefore are more penetrating, detection of these electron-like particles mainly relies on sensing the free ionization they produced by interacting with ambient atoms in the detecting medium.

Fission fragments and products Fission fragments and fission products are often confused, but there is an important difference. Fragments are the particles released immediately after a fission (scission) occurs, usually resulting in the release of two, but sometimes three, fission fragments. These highly excited and highly ionized fragments of “nuclear fluid” are extremely unstable. They emit almost all fission neutrons within 10 14 s of the fission event and, through isomeric decay, the fragments emit several MeV of prompt gamma rays within the first 60 ns after a fission. For example, thermal-neutron induced fission of 92235U with a Q-value of 202.5 MeV yields 8.13  0.35 prompt gamma rays having an average total energy of 7.25  0.26 MeV per fission, distributed over a continuum of energies between 0.1 MeV and 10.5 MeV (Shultis and Faw, 2000). On average, prompt neutrons have about 4.9 MeV, while the combined fission fragments have about 169 MeV. The remaining energy is divided between b, neutrino, and delayed g radiation. After promptneutron and gamma-ray emission, the still highly ionized fast-moving fragments rapidly slow because of their high collisional stopping power –dE/dx. During the slowing process they capture ambient electrons to become neutral fission products which then undergo b decay over the next several hundred years through a chain of radioactive daughters. Notably, there are seven fission products with half-lives greater than 105 years. As with other charged particles, ionization within the detector medium is commonly used for fission fragment detection, although other detection methods do exist.

Physical Sensors: Radiation Sensors

143

Electromagnetic radiation Different types of ionizing electromagnetic radiation are defined by the mechanism of its origin and not by the energy of its photons. The energetic photons of the different types of electromagnetic radiation are generally detected by similar means regardless of their origin. The principal interactions used for photon detection are the photoelectric effect, the Compton scattering effect, and the pair production effect. These effects produce, either directly or indirectly, energetic charged particles, which, in turn, create additional charge carriers through Coulombic interactions with the ambient medium. Usually photons are detected by sensing the presence of the charge carriers, although other techniques can be used.

X-rays X-rays were discovered by Wilhelm Röntgen in 1895 while he was conducting experiments with a Crookes tube (Röntgen, 1895, 1896), a type of cathode-ray tube. While in a darkened room, he noticed a plate, coated with a scintillating mixture of platinobarium-cyanide (BaPt(CN)4), would glow when the Crookes tube was activated. This led to his conclusion that invisible rays were emanating from the Crookes tube and were causing the glowing effect. Although X-ray photons generally have lower energies than those of most gamma rays, it is incorrect to differentiate between Xrays and gamma rays by their energy. X-rays are ionizing photons produced when a charged particle, such as an electron, undergoes a reduction in energy either by electronic transitions from an atomic orbital to a lower energy orbital or by interactions with a strong electromagnetic field. By contrast, gamma rays are released by the transitions of a nucleon from an excited state in the nucleus to a lower energy state. Characteristic X-rays are released with discrete energies by electronic transitions from higher to lower atomic orbital energy states. They are categorized by the initial and final electronic shells (orbitals). For instance, an electron dropping from the L shell to the K shell yields a Ka X-ray and from the M shell to the K shell yields a Kb X-ray. Transitions from an outer shell down to the L shell are L X-rays, where M to L yields an La X-ray and a N to L generates an Lb X-ray. Bremsstrahlung X-rays (“breaking radiation”) are emitted over a continuous spectrum as fast electrons (or beta particles) centripetally decelerate in the strong electric field near the nucleus of an ambient atom. The endpoint energy of the emitted continuous photon spectrum is the initial energy of the electron when it enters the slowing medium. However, the most probable X-ray energy is usually substantially less than this maximum energy. It is bremsstrahlung radiation, usually accompanied by characteristic X-rays, that is emitted from common X-ray tubes. Bremsstrahlung can also be emitted by any charged particle, but electron generated bremsstrahlung is usually of most concern. Synchrotron/Cyclotron X-rays, also known as magnetobremsstrahlung, are produced by accelerating charged particles, usually from interactions with a strong magnetic field such as those encountered in particle accelerators or active astronomical objects such as pulsars and black holes.

Gamma rays Gamma rays (g rays) were discovered by Paul Villard in 1900 while conducting experiments with a radium source and photographic film (Villard, 1900a, 1900b). Within a strong magnetic field, he noticed that three visible areas were produced on a photographic plate. One radiation type was easily blocked by a thin shield and was also deflected as positive particles (a particles), one type was deflected as negative particles (b particles), and a third type was not deflected at all and was also highly penetrating. Villard did not name this newly discovered radiation, but instead it was Ernest Rutherford who offered the name “g radiation” in sequence with a and b radiation. A gamma ray is electromagnetic ionizing radiation emitted from an atomic nucleus transitioning from an excited state to a lower energy state. Gamma rays can be highly penetrating and their detection is improved by using detectors constructed from relatively high atomic number materials and large detector volumes.

Annihilation photons Annihilation photons are also identified by their origin, namely, the destructive joining of a subatomic particle to its antiparticle. For practical detection purposes, the most common annihilation photons encountered are 511-keV photons emitted from the combination of an electron with a positron. The 511-keV energy is the rest mass energy equivalent of an electron (or positron). Usually annihilation occurs after an initially fast moving bþ particle has completely slowed before annihilating with an ambient electron so, consequently, the two 511-keV annihilation photons are emitted in opposite directions to preserve the zero linear momentum condition. However, so-called in-flight annihilation does occur. There are certainly more complicated annihilation schemes, such as found with baryon-antibaryon reactions and neutrino-antineutrino reactions, yet these reactions can be quite involved and generally beyond the scope of the present article on radiation detection.

Neutral particles Neutrons Neutrons were discovered by James Chadwick in 1932, who received the 1935 Nobel Prize in Physics for this discovery (Chadwick, 1932). Neutrons are subatomic particles with zero charge and mass of 1.674927  10 27 kg, slightly more than that of a proton. A neutron is composed of two down quarks and one up quark. Neutrons are fundamental baryon particles that make up atomic

144

Physical Sensors: Radiation Sensors

nuclei, the others being protons (1.672621  10 27 kg). Although neutrons have no electronic charge, they can still cause ionization in a medium through absorption or scattering reactions.

Neutrinos and anti-neutrinos A neutrino is a subatomic particle with mass much smaller than other elementary particles. It has spin of 1/2 with a net neutral charge. They are weakly interacting particles and usually pass through material without interacting. The anti-neutrino is similar to the neutrino, also neutral, but with opposite lepton number and chirality. There are three types of neutrinos: the electron neutrino, the muon neutrino, and the tauon neutrino which oscillate from one form to another as they travel. Electron antineutrinos are emitted in beta-particle radioactive decay. Massive detectors are required for neutrino detection, often taking advantage of the Cerenkov radiation effect in transparent materials (water, heavy water, etc.).

Gas-filled detectors In 1908, Ernest Rutherford and Hans Geiger constructed a device composed of a metallic cylinder containing a thin wire along the axis (Rutherford and Geiger, 1908). The gas medium in the device was air. Application of a voltage between the cylinder and wire produced sizable currents, measured with an electrometer, when a particles were directed into the cylinder. They also noticed that the behavior of the detector changed with increasing voltage. In particular, a particles could be detected at much lower applied voltages than could be b particles. This behavior and its application is referred to as proportional counting. Several distinct voltage regions of operation are clearly apparent in Fig. 1. Direct radiation interactions in the gas or particles ejected from the chamber walls by radiation interactions cause some of the detector gas to become ionized. Through a cascade of ionization from primary and secondary ionization, a charge cloud composed of electrons and positive ions appears. A voltage placed across electrodes in the gas chamber causes the electrons and ions to drift apart, where electrons drift towards the anode and the positive ions drift towards the cathode. As the charge carriers move through the chamber, they induce current to now in a circuit externally connected to the chamber. This current, or change in current, can then be measured as an indication that a radiation interaction occurred in the chamber (Korff, 1946; Wilkinson, 1950).

General operation At low voltages in the recombination region, denoted as Region I of Fig. 1, some measurable current is seen although considerable recombination occurs. As the voltage is increased, electron-ion pair separation becomes more efficient until practically no recombination occurs. Hence, the measured current is indicative of the total number of electron-ion pairs formed. Negligible recombination occurs in Region II of Fig. 1 and is referred to as the ionization chamber region. As the voltage is increased further, the electrons gain enough kinetic energy to create more electron-ion pairs through impact ionization. This ionization provides a mechanism for signal gain, often referred to as gas multiplication. The observed current increases as the voltage increases, but still remains proportional to the energy of the original radiation particle. This voltage region is shown on the gas curve as Region III. Increasing the ion pair recombination occurs before collection

Geiger-Muller region

Pulse Height or Ions Collected (log scale)

onset of “gas multiplication” transition region

I

II

III particles

IIIa

IV

V

particles

Detector High Voltage (volts) Fig. 1 The observed output pulse height versus the applied high voltage for a gas-filled detector, showing the main regions: (I) recombination, (II) ion chamber, (III) proportional, (IV) Geiger-Müller, and (V) continuous discharge. This plot is often referred to as the gas curve. Copyright (2020). From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

Physical Sensors: Radiation Sensors

145

applied voltage further causes nonproportional current increases and is indicated as Region IIIa. At even higher electric fields, defined as region IV, all currents, regardless of origin, radiation species, or energies, have the same magnitude. This region is called the Geiger-Müller region. Finally, excessive voltage drives the detector into Region V where the voltage causes sporadic arcing and other spontaneous electron emissions to occur, hence causing continuous discharging in the detector. Gas detectors should not be operated in the continuous discharge region. Gas detectors can be operated in pulse mode or current mode. Pulse mode is generally used in low to moderate radiation fields. For pulse mode, a single radiation quantum, such as an alpha particle, beta particle or gamma ray, interacts in the chamber volume and produces an ionized cloud. The charge carriers drift apart and they induce current to flow to the device terminals. A charging circuit, usually consisting of a preamplifier and feedback loop, integrates the current and stores the charge, thereby producing a voltage potential. This voltage is measured as a single event, indicating that a single radiation quantum has been detected. The preamplifier circuit is subsequently discharged and reset, allowing the device to measure the next radiation interaction event. Hence, each voltage pulse from the detector indicates an individual radiation interaction event. Although extremely useful, there are limitations to pulse mode operation. Should another radiation interaction occur while the detector is integrating or discharging the current from a previous interaction event, the device may not, and usually does not, record the new interaction, a condition referred to as pulse pile up. The time duration in which a new pulse cannot be recorded is the detector recovery time, sometimes referred to as dead time. For high radiation fields, gas detectors are operated in current mode, in which the radiation induced current is measured on a current meter. Under such conditions, many interactions can occur in the device in short periods of time, and the observed current increases as the radiation interaction rate in the chamber increases. Hence, current mode can be used to measure the intensity of high radiation fields; the magnitude of the current is a measure of the radiation-induced ionization rate in the detector, which, in turn, is a measure of the strength or intensity of the radiation field in which the device is being operated. The disadvantage of current mode operation is that individual radiation interactions cannot be identified.

Ion chambers Ion chambers are operated in Region II of the gas curve (Fig. 1). Ion chambers come in many forms, and can be used for reactor power measurements, where the radiation field is very high, survey instruments in moderate radiation fields, and as small personnel dosimeters, for use where radiation levels are usually low. When ionizing radiation interactions occur in the gas, they produce electron-ion pairs relative in number to the radiation energy absorbed. The voltage applied across the electrodes causes the negative electrons to separate from the positive ions and drift across the chamber volume. Electrons drift towards the anode and positive ions drift towards the cathode, and their movement induces current to flow in the external circuit. Two difficulties are encountered when operating ion chambers in the pulse mode: (1) the measured signal is small because the current is composed of only the primary (or initial) electron-ion pairs excited by the radiation quantum and (2) the signal formation time can be long due to the slow motion of the heavy positive ions. Often, an RC circuit is connected to an ion chamber to reduce the time constant of the system and to discharge the capacitor before all of the ions are collected in order to reduce response time. Ion chambers operated in the current mode are stable and have a long life. Large ion chambers are used as area monitors for ionizing radiation and high-pressure chambers offer a relatively high sensitivity, permitting measurement of exposure rates as low as 1 mR/h. At the other extreme, small chambers with low gas pressures can be operated in radiation fields with exposure rates as great as 107 R/h. An ion chamber can be coated with a strongly-absorbing neutron-reactive material or filled with a neutron reactive gas, thereby making it sensitive to neutrons (Grosshoeg, 1979). Common isotopes for gas-filled neutron detectors are 3He, 10B, 6Li, and 235U. Neutron sensitive ion chambers can be filled with 10BF3 or 3He gas, or the inside walls of the chamber are coated with 10B, 6LiF, or 235 U. Ion chambers that use 235U are often referred to as fission chambers, because it is the fission fragments from the fission of 235U that ionize the chamber gas. These gas-filled neutron detectors can be operated as ion chambers or proportional counters. Because of pulse pile up, ion chambers and fission chambers are generally not operated in pulse mode when in high radiation fields.

Proportional counters At sufficiently high voltages, electrons can gain enough kinetic energy to cause additional ionization and excitation in the gas, an effect called impact ionization. These newly liberated electrons can also gain enough energy when accelerated by the electric field to cause even more ionization. The process continues until the electrons are collected at the anode. The entire process of generating the impact ionization cloud is called a Townsend avalanche, or sometimes gas multiplication, as illustrated in Fig. 2. There is a critical electric field EA at which gas multiplication begins and below which the electrons do not gain sufficient energy to cause impact ionization. This threshold electric field defines the transition between Region II and Region III in the gas curve. Detector configurations of parallel plates, which can be used for ion chambers, are seldom used for proportional counters. A preferred geometry is either a coaxial configuration (depicted in Figs. 2 and 3) or a pseudo-hemispherical geometry. Pulse formation is dependent on the formation of a large electron-ion charge cloud near the anode. Consequently, it is the drift of the positive ions that induces most of the output signal. Approximately half of the maximum signal is formed as positive ions drift within the first mm of the anode wire towards to the cathode. The dead time is reduced by adjusting the RC time constant of the device with only minor effect on the overall performance. Dead times for typical proportional counters are on the order of tens of microseconds.

146

Physical Sensors: Radiation Sensors ionizing radiation

thin anode wire

esignal high field region eTownsend avalanche

ionizing radiation

V

Fig. 2 With a high electric field near the anode of a gas-filled detector, signal gain is realized through impact (Townsend) avalanching and is referred to as “gas multiplication”. Copyright (2020). From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

Electron-Ion Pairs

CUT AWAY VIEW Thin Window

Insulator

Ionizing Radiation

Anode Wire Gas Container and Cathode Fig. 3 Schematic view of a coaxial gas detector, which is commonly used for Geiger-Müller tubes, and sometimes used for proportional counters. High voltage is applied to the central wire anode, while the outer cylinder wall, the cathode, is held at ground. Copyright (2020). From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

Continuous waves of avalanches can occur if ultraviolet (UV) light, released by the excited atoms or ions produced in one avalanche, ionize too many gas atoms after the initial ionization event. These gas ions, when arriving at the cathode wall, can strike with enough kinetic energy to cause the ejection of more electrons, causing another avalanche (Liebson, 1947; Sitar et al., 1993). To prevent continuous waves of avalanching from occurring in the chamber after a radiation interaction, a quenching gas is added to the gas mixture, typically a polyorganic molecule. A common proportional counter gas is P-10, which is a mixture of 90% Ar and 10% methane (the quenching gas). When an ionizing particle enters the detector, it ionizes both the Ar and the quenching gas. However, as the Ar gas ions drift through the chamber, they transfer their charge to the quench gas molecules. The quench gas ions continue to drift and carry the positive charge to the cathode wall. When a quench gas is struck by a UV photon or strikes the cathode wall, it dissociates by releasing a hydrogen atom rather than ejecting an electron. Hence, the quench gas prevents continuous waves of avalanches. As with ion chambers, a neutron detector can be fashioned with a proportional counter by backfilling the chamber with a neutron reactive gas (usually 3He or 10BF3). Because such a device operates in proportional mode, a low resolution spectrum associated with the reaction product energies of either the 10B(n,a)7Li reaction or the 3He(n,p)3H reaction can be identified, depending on the gas used in the counter. Neutron detectors are also fabricated by coating the walls with a neutron reactive material (usually 10 B or 10BC4) with the chamber backfilled with a common proportional-counter gas.

Geiger-Müller counters Although Hans Geiger originally created the gas-filled detector in 1908 (with Ernest Rutherford), the device used today is based on an improved version that Geiger’s first PhD student, Walther Müller, constructed in 1928 (Geiger and Müller, 1928). Hence, the proper name for the device is the “Geiger-Müller” counter. The original “Geiger” counter was sensitive to alpha particles, but not

Physical Sensors: Radiation Sensors

147

so much to other forms of ionizing radiation. Müller’s improvements included the implementation of vacuum tube technology, which allowed the device to be formed into a compact and portable tube sensitive to alpha, beta, and gamma radiation. In 1947, Sidney Liebson further improved the device by substituting a halogen as the quenching gas (Liebson and Friedman, 1948). This change in the quenching gas enabled the detector to operate at lower applied voltages over a longer lifespan. Geiger-Müller (GM) counters are typically arranged in a coaxial configuration, in which a thin anode wire is projected inside a tube that serves as the cathode. A high voltage is applied to the central anode wire, while the cathode is held at ground. A GM counter is operated in Region IV of the gas curve and depends upon gas multiplication as a signal amplification mechanism, much like a proportional counter does. However a single important difference is that, at any specific applied voltage, all output pulses from a GM counter are of the same amplitude regardless of the ionizing radiation energy or type. This constant amplitude results from how the pulses are formed. In a proportional counter the charge cloud formation ends when the free electrons are collected. But in a GM counter the charge cloud formation ends when the space-charge buildup reduces the electric field below EA. Because of the constant pulse amplitudes, GM counters do not intrinsically possess the ability to discriminate among alpha, beta, or gamma radiations, nor can they distinguish the different energies of these radiations. Pulse formation time is relatively long for a GM counter, and consequently, the dead time is typically 10 times longer than that of proportional counters of similar size and is typically about 100 ms. Because GM counters are usually sealed tubes, the quenching gas inside can be exhausted over time if traditional organic molecules such as the methane component of P-10 gas are used. Instead, Geiger-Müller counters use halogens for a quenching gas, in which the diatomic molecules dissociate when they strike the cathode. Halogens, unlike methane, can heal themselves by recombining into diatomic molecules, thereby extending the life of the gas in the detector.

Scintillation detectors Scintillation detectors convert absorbed radiation energy into detectable light photons, usually in the visible light range. These light emissions can then be detected with light-sensitive instrumentation. Scintillators are generally separated into two classes: inorganic and organic scintillators. This distinction is made because the mechanisms of light production are different for scintillators in each category.

Inorganic scintillators Inorganic scintillators can be found as crystalline, polycrystalline, or microcrystalline materials and depend primarily on the energy band structure of the material for the production of scintillation light (McGregor, 2018). Depicted in Fig. 4 are the energy bands of inorganic scintillators. If a radiation particle interacts in the scintillator, it can excite, through a cascade of processes, numerous electrons from the valence and lower-bound bands up into the conduction bands (see Fig. 4A) or the exciton band. An exciton is an electron and hole bound together as a neutral quasi-particle. These electrons rapidly lose energy and fall back to the conduction band edge Ec. As electrons and excitons deexcite and drop back into the valence band, they can lose energy through light emissions. However, if the emitted photons have an energy equivalent to the band gap energy Eg, the emitted photons can be reabsorbed by the scintillator. In effect, the scintillator is opaque to its own emissions. To eliminate this opaqueness, an impurity is added to the crystal that introduces energy states in the band gap (depicted in Fig. 4B) producing an activated scintillator (Birks, 1964). In an activated scintillator, the excitation process is the same as in a crystal without the activator but now a significant fraction of excited electrons can fall into the activator excited state. Subsequent transitions to the ground state release photons with sub-band-gap energies and, thus, avoid reabsorption. However, there are exceptions in which intrinsic scintillators (scintillators without an activation impurity) still work well such as bismuth germanate (Bi4Ge3O12 or BGO) and barium fluoride (BaF2) (Rodnyi, 1997). The light emission spectrum is dependent on the activator choice and is usually spread over a continuum of energies (Lecoq et al., 2006). The configuration coordinate diagram of Fig. 5 can be used to explain the emission spectrum (Ridley, 1982; Yen et al., 2007). An electron can be excited to point B by excitation from point A, can be captured from the conduction band, or can experience exciton capture. The transition causes thermal motion and vibrational states to appear that are represented as horizontal lines in Fig. 5. Both the spatial occupancy and the electron wave function change with this transition to the excited state, and the spatial position of lowest energy changes from A to QE0 (Ridley, 1982). The electron rapidly loses energy and moves to location C and can drop to location E, releasing a photon. The electron thermally loses energy and returns to location A. This change in the most probable absorption wavelength to the most probable emission wavelength is called the Stokes Shift. Scintillator activator impurities are identified by separating them by a colon or by parentheses. For example, Tl is the activator in NaI:Tl or NaI(Tl). The intensity of light emitted at time t after radiation is absorbed in the scintillator is I(t) ¼ I(0) exp.( t/s), where s is the mean decay time. The brightness or light yield of a scintillator is a measure of photon yield per unit energy, with higher light yields usually producing better energy resolution. Discovered in 1948 by Robert Hofstadter, NaI:Tl is one of the most widely used scintillators for gamma-ray spectroscopy (Hofstadter, 1949). NaI:Tl detectors do not require cooling during operation and can be used in a great variety of applications. The bare NaI:Tl crystal is hygroscopic and fragile; however, when properly packaged, they can have a long life, operate in warm and humid

148

Physical Sensors: Radiation Sensors

(A)

(B)

Upper Band

Forbidden Band

Conduction Band

Ec

Band Gap

Eg

Exciton

Exciton Band

Eg

Et1a Et1b

Exciton

Exciton Band

Et0

Ev Valence Band Forbidden Band Tightly Bound Band

Fig. 4 Shown are two basic methods by which an inorganic scintillator produces light. The intrinsic case (A) has no added activator impurities. The extrinsic case (B) has activator impurities that produce energy levels in the band gap. Copyright (2020). From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

E

excited

ground

D

D‛ E2

B h

C

in

h

E1

out

Intensity (relative units)

vibrational states (Stokes Shift)

excitation

emission

E0

E vibrational states

A

Qe0

Q

max1

max2

Wavelength

Fig. 5 Configuration coordinate diagram for a luminescent center depicting luminescence with a Stokes shift and also a form of non-radiative deexcitation. The ordinate is energy E and the abscissa is the average distance Q between the luminescent center and the surrounding ions. Copyright (2020). From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

Physical Sensors: Radiation Sensors

149

environments, resist a reasonable level of mechanical shock, and are resistant to radiation damage. For common applications requiring high gamma-ray efficiency and a modest energy resolution, NaI:Tl is a good choice (Shafroth, 1967). Since the discovery of NaI:Tl, the search has continued for a scintillator with better performance for gamma-ray detection and spectroscopy. A detailed discussion of the search for new scintillators is provided by Lecoq et al. [2006] (see also McGregor and Shultis, 2020, and references therein). In recent years, LaBr3:Ce, a relatively new scintillator with exceptional properties for gamma-ray spectroscopy, has become available. LaBr3:Ce has a much higher light yield and a much shorter decay time than those of NaI:Tl. Cs2LiLaBr6:Ce (CLLB:Ce) and Cs2LiYCl6:Ce (CLYC:Ce) both have good gamma-ray spectroscopic performance with the added benefit of being neutron sensitive from the 6Li(n,t)4He reaction. Bi4Ge3Q12 (BGO) has a much better detection efficiency than NaI:Tl, but somewhat inferior energy resolution. A comparison of spectroscopic performance between three commercially available scintillators is shown in Fig. 6.

Organic scintillators Organic scintillators depend primarily on the molecular structure of the material for the scintillation mechanism (Birks, 1964). A typical Jablonski energy diagram is shown in Fig. 7 for an organic scintillator. An independent molecule of organic scintillation material can have an electron excited through the p states up from the ground state into an excited singlet state, of which there are many levels. There are many vibrational levels associated with the ground states, typically denoted by S0X where x refers to one of the vibrational sub-states. There are also numerous excited singlet states as well as excited triplet states associated with the carbon p bonds. Electrons that gain energy rise to one of the excited vibrational states and generally fall rapidly to the lowest S10 state, which then deexcite through two possible channels. If the electrons deexcite directly from the S10 state to one of the S0X states, the light emission is rapid and is referred to as scintillation fluorescence. Decay times for fluorescence are typically only a few nanoseconds, and fluorescent emission can be easily linked to individual radiation events. However, if the electrons deexcite by crossing to the triplet states T1X and then eventually fall to one of the S0X states, the light emission is slow and is referred to as scintillation phosphorescence (Birks, 1964). This second light producing mechanism is undesirable because phosphorescent emission is slow and continues to produce afterglow for extended periods of time and, hence, cannot be directly linked to individual radiation events, especially in high-radiation fields. Organic scintillators are composed mostly of hydrogen and carbon, both of which are poor absorbers of gamma rays. Organic scintillators have nonlinear light output for heavy ion radiation, yet much more linear in response to electrons and beta particles. Organic scintillators are frequently used for beta particle and electron detection (NCRP, 1985). Because much of the energy of fast neutrons, upon scattering from H, is transferred to the recoil proton, organic scintillators are often used for fast neutron detection by detecting the recoil protons. Organic scintillators depend on the molecular structure, often a benzene ring structure, and do not need activator dopants for the scintillation mechanism. Hence, they also need not be crystalline or polycrystalline in structure. As a result, organic scintillators can be formed as solids, liquids, gases, and plastics. Liquid scintillators are commonly used for beta particle measurements, where the radioactive material is mixed in a “cocktail” solution of fluor and solvent (Horrocks and Peng, 1971; NCRP, 1985). Small amounts of materials can be mixed in the organic scintillator without destroying the scintillation process. For instance 10B, 6LiF, or Gd can be

Fig. 6 Comparison of normalized spectral performance for a 2 inch  2 inch NaI:Tl detector, a 2 inch  2 inch LaBr3:Ce detector, and a BGO detector of similar size for 662-keV gamma rays from 137Cs. Copyright (2020). From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

150

Physical Sensors: Radiation Sensors

-states

I singlet states S30

S2

S21 S20

S1

S13 S12 S11 S10

T3 inter-system crossing

S3

triplet states

T2

T1

S0

S03 S02 S01 S00

fluorescent emission Fig. 7

phosphorescent emission

Shown are two basic methods by which an organic scintillator produces light. After Birks [1964].

mixed into the scintillator to enhance thermal neutron sensitivity while heavy metal particles, such as Pb, can be added to increase gamma-ray sensitivity.

Light collection Although a scintillator produces a light signal when a radiation particle interacts in the scintillator, the light must be converted to an electronic signal if it is to be recorded. Below, four devices for doing such a conversion are discussed.

Photomultiplier tubes Secondary electron emission was discovered by Austin and Starke (1902) when they noticed that exposing metal surfaces to cathode rays (electrons) caused the emission of more electrons than were incident. In 1919, Slepian proposed, and patented, the concept of using secondary electron emission as an amplification device (Slepian, 1923). Some attempts were made to use the process with vacuum tube technology, but it was not until 1941 that the RCA Company released the first amplifier tube, called a photomultiplier tube (PMT), that used secondary electron emission for signal amplification (Engstrom et al., 1980). This PMT, the RCA Type 931, was originally used as a signal amplifier for electronics, and was used for radar jamming technology during the Second World War. In 1947 von Broser and Kallman (1947a, 1947b) coupled a PMT to the organic scintillator naphthalene and produced the first scintillation detector system. In 1948 Hofstadter discovered the inorganic scintillator NaI:Tl, which when coupled with a PMT produced the first practical solid-state gamma-ray spectrometer. The basic PMT, depicted in Fig. 8, has a photocathode to absorb light emissions from a light source such as a scintillating material (see Engstrom et al., 1980). Light photons that strike the photocathode excite electrons that can then diffuse to the surface facing the vacuum side. A fraction of these electrons escape the surface into the vacuum tube. A voltage applied to the tube accelerates and guides these electrons to the first dynode electrode. When an electron strikes the dynode, it again causes more electrons to become liberated into the tube. These newly liberated electrons are then guided to the next dynode where more electrons are liberated and so on. As a result, the total number of electrons eventually released depends on the number of dynodes in the PMT and the photoefficiency of the photocathode and emission efficiency of the dynodes. The total charge released in the PMT is Q ¼ qN0GN, where q is the charge of an electron, N0 is the initial number of electrons released at the photocathode and that reach the first dynode, G is the number of electrons released per dynode per incident electron (the gain), and n is the total number of dynodes in the PMT. For instance suppose that a PMT has 10 dynodes each operated with a gain of 4. An event that initially releases 1000 electrons (N0) causes over 109 electrons to emerge from the PMT.

Physical Sensors: Radiation Sensors detector window

151

photocathode dynodes

scintillator anode

-ray

current output

reflector scintillation coupling compound photons

photoelectron PMT window

vacuum tube

Fig. 8 The basic mechanisms of a photomultiplier tube (PMT). An absorbed g ray causes the emission of numerous light photons, some of which reach the photocathode. Electrons emitted from the photocathode are then amplified in number by a chain of dynodes. Copyright (2017). From Fundamentals of Nuclear Science and Engineering by J.K. Shultis and R.E. Faw. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

PMTs are stable and electronically quiet (low noise). Modern varieties have high photocathode and dynode efficiencies, often referred to as quantum efficiency, with gains that can exceed 30. Photocathodes are generally made from alkaline metals, which are most sensitive to light in the 350 to 450 nm range (Donati, 2000). Scintillators emitting light outside this range can still be used under some circumstances, although their effectiveness can be compromised.

MicroChannel plates Microchannel plates are an alternative method for amplifying signals from a scintillator. Microchannel plates are glass tubes whose insides are coated with secondary-electron emissive materials. A voltage is applied across the tube length which causes electrons to cascade down the tube. Every time an electron strikes the tube wall, more electrons are emitted, much like with dynodes in a PMT. Hence, a single electron entering the tube can cause a cascade that can eventually produce 106 electrons emitted from the other end of the tube (Donati, 2000). Typically, hundreds of these microchannels are bonded together to form a plate of channels running in parallel. The microchannel plate can be fastened to a scintillator to operate in a similar fashion as a PMT. Light photons entering the microchannel plate cause the ejection of primary photoelectrons, which cascade down the microchannels to liberate millions more electrons. The main advantage of a microchannel plate over a PMT is its compact size. A microchannel plate only 1 inch thick can produce a signal of similar strength as a PMT. The main problem with microchannel plates is the signal produced per monoenergetic radiation event is statistically much noisier than that produced by a PMT, hence the energy resolution for spectroscopy is typically worse than that provided by a PMT.

Photodiodes Photodiodes are semiconductor devices formed into a pn or pin junction diode (Sze, 1981). When photons strike the semiconductor, usually Si or GaAs based materials, electrons are excited. A voltage bias across the diode causes the electrons to drift across the device and induce charge much as occurs in a gas-filled ion chamber. The quantum efficiency of the semiconductor diode varies with the configuration and packaging of the diode. For instance, various different commercial Si photodiodes have peak efficiencies at wavelengths ranging between 700 and 1000 nm (Sze, 1981; Donati, 2000). Regardless, they are typically more sensitive to longer wavelengths than commercial PMTs. As a result, emissions from CsI:Tl, a relatively bright scintillator with most probable emissions near 550 nm, match better to Si photodiodes than PMTs. Photodiodes operate with low voltage, are small, rugged, and relatively inexpensive. Thus, they provide a practical and compact method of sensing light emissions from scintillators. However, they typically do not couple well to light emissions near the 400 nm range (blue-green) and have low gain, if any at all. Consequently, the signals from photodiodes need more amplification than signals from PMTs, and scintillator/photodiode systems generally do not have an energy resolution as good as that of scintillator/PMT systems.

Silicon photomultipliers A single photon avalanche diode (SPAD) can be used to increase the signal gain (Cova et al., 1989). The device is designed such that an electric field in the diode is sufficiently high to cause impact ionization of more electrons, in a similar fashion as occurs in proportional and Geiger-Müller counters. If the device is overly biased, beyond stability, then a single photon can trigger an avalanche (Haitz, 1961; McIntyre, 1961). The avalanche can be quenched by including a resistor in series with the diode. As the current increases, the voltage across the resistor also increases, thereby reducing the SPAD voltage below the critical electric field strength and, thereby quenching the avalanche.

152

Physical Sensors: Radiation Sensors

An array of SPADs can be fashioned into a type of silicon photomultiplier (SiPM) (Renker, 2006). The device operates under the assumption that, statistically, a single SPAD will intercept no more than one photon per scintillation shower. Each SPAD is designed and biased such that they all produce approximately the same signal output, much like the GM counter. Consequently, the total output voltage can be divided by the average voltage per SPAD to determine the total number of photons in the shower. Nonlinearities can arise from high-energy gamma rays interacting in exceptionally bright scintillators such that more than one photon strikes an SPAD simultaneously. The consequence is that the SPAD cannot distinguish between one or more photons, and the SiPM output voltage will underestimate the true number of photons and the associated gamma-ray energy. SiPMs are compact and rugged devices, and are generally unaffected by magnetic fields. In recent years, SiPM technology has become a popular alternative to PMTs for commercial scintillation detectors.

Semiconductor detectors Perhaps the first reported attempt of producing a solid-state conduction counter is by Röntgen and Joffé [1913] on various solids, yielding mostly unsuccessful to marginal results. The first successful semiconductor detectors were reported by Stetter (1941) (with diamond samples) and then by Van Heerden (1945) (AgCl samples). Although afterwards various other semiconductors were developed and tested for novel radiation detection applications, it was not until Pell introduced the Li drifting compensation method in 1960 did certain semiconductors (mainly Si and Ge) become important radiation detectors. The operation of a semiconductor detector combines the charge excitation process in a crystalline inorganic scintillator and the charge collection method of a gas-filled ion chamber. Shown in Fig. 9, gamma rays or charged particles absorbed in the semiconductor excite electrons from the valence and tightly bound energy bands up into the numerous (and almost continuum of) conduction bands. The empty states left behind by the negative electrons act as positively charged particles, called holes. The excited electrons rapidly deexcite to the conduction band edge EC. Likewise, as electrons high in the valence band fall to lower empty states in the valence and tightly bound bands, the holes appear to move upwards towards the valence band edge EV.

Upper Band

Forbidden Band

negative voltage reference

Conduction Band

Ec Band Gap

Ev

+

“electrons”

Eg “holes”

positive voltage reference

Valence Band Forbidden Band Tightly Bound Band

-ray Fig. 9 Absorbed radiation energy excites electrons from the valence and tightly bound bands up into the higher conduction bands. The empty states (holes) left behind mimic positive charge carriers. The electrons quickly deexcite to the lowest conduction band edge EC and the holes rapidly deexcite to the top of the valence band EV. A voltage applied to the detector causes the electron and hole charge carriers to drift to the electrodes. Copyright (2017). From Nuclear Science and Engineering by J.K. Shultis and R.E. Faw. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

Physical Sensors: Radiation Sensors

153

The principal difference between a semiconductor and almost all scintillators is that the mobility of charge carriers in semiconductors is sufficiently high to allow conduction, whereas scintillation materials are mostly insulating materials that do not conduct. As a result, a voltage can be applied across a semiconductor material to cause the negative electrons and positive holes, commonly referred to as electron-hole pairs, to drift in opposite directions, much like the electron-ion pairs do in a gas-filled ion chamber. In fact, at one time semiconductor detectors were referred to as “solid-state ion chambers” (Price, 1964). As these charges drift across the semiconductor, they induce a current to flow in an external circuit which can either be measured as a current or stored across a capacitor to form a voltage. Semiconductors are far more desirable for energy spectroscopy than gas-filled detectors or scintillation detectors because they are capable of much higher energy resolution. This observed resolution improvement is largely due to the better statistics resulting from the more numerous charge carriers produced by a radiation interaction (Bertolini and Coche, 1968). Typically, it only takes between 3 eV and 5 eV to produce an electron-hole pair in a semiconductor (McGregor and Hermon, 1997; Owens, 2019), whereas 25– 35 eV are required to produce an ion-electron pair in most gases (Sauli, 2014). Some common semiconductors and their properties are listed in Table 1. Most semiconductor detectors are configured as either planar or coaxial devices, as depicted in Fig. 10. Small semiconductor detectors are configured as planar devices and can be used for charged-particle detection and gamma-ray detection (Dearnaley and Northrop, 1966). Large semiconductor gamma-ray spectrometers are usually configured in a coaxial geometry to reduce the detector capacitance. Refrigeration with cryogenics or mechanical coolers are used with many large semiconductors, usually Ge and Si devices, to reduce thermally generated leakage currents. To reduce injected leakage current, semiconductors are formed as reverse biased diodes, which can be Schottky, pn, or pin junction devices. Under some special conditions, highly resistive semiconductors need only have ohmic contacts because the bulk resistance is adequate to reduce leakage currents (McGregor and Shultis, 2020).

Ge detectors In 1960, Pell introduced the technique for drifting Li ions into a silicon semiconductor to counteract the deleterious effects of impurities (Pell, 1960). Within a few years, the same technique was adopted for Ge detectors (Brown et al., 1969). Unfortunately, if Table 1

Common semiconductors and their properties.

Semiconductor

At. Number (Z)

Density (g cm 3)

Band Gap (eV)

Ionization Energy (eV/e-h pair)

Si Ge GaAs CdTe Gd0.9Zn0.1Te HgI2

14 32 31/33 48/52 48/30/52 80/53

2.33 5.33 5.32 6.06 6.0 6.4

1.12 0.72 1.42 1.52 1.60 2.13

3.61 2.98 4.2 4.43 5.0 4.3

semiconductor

semiconductor

-ray holes

electrical contact

+ ++ +

++ + +

--

- --

electrons

holes electrons

electrical contact

Planar Design

+

electrical contact

-ray

+

electrical contact

Coaxial Design (cross section)

Fig. 10 The two most common designs for semiconductor detectors are the planar and coaxial configurations. Copyright (2017). From Fundamentals of Nuclear Science and Engineering by J.K. Shultis and R.E. Faw. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

154

Physical Sensors: Radiation Sensors

Fig. 11 Comparison of the energy resolutions of a NaI:Tl and an HPGe detector. The gamma-ray source is a mixture of the radioisotopes 152Eu, 154 Eu, and 155Eu. Copyright (2017). From Fundamentals of Nuclear Science and Engineering by J.K. Shultis and R.E. Faw. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

a lithium-drifted Ge [or Ge(Li)] detector warms to room temperature, the Li diffuses from compensation sites and the detector is ruined (Bertolini and Coche, 1968). Consequently, Ge(Li) detectors must always be keep at liquid nitrogen (LN2) temperatures. Modern Ge detectors are made from high-purity zone-refined material and no longer require Li compensation. However, a high purity Ge (HPGe) detector must still be chilled with LN2 or a refrigerator when operated in order to reduce thermally generated leakage currents. HPGe detectors have exceptional energy resolution compared to scintillation and gas-filled detectors. Comparison spectra between a NaI:Tl and a HPGe spectrometer are shown in Fig. 11. HPGe detectors are now standard high-resolution spectroscopy devices used in the laboratory. Their high energy resolution allows them to easily identify radioactive isotopes for a variety of applications such as impurity analysis, composition analysis, and medical isotope characterization, to name a few. Portable devices with small LN2 dewars or compact efficient refrigerators are also available for remote spectroscopy measurements. Because of the higher atomic number of iodine and the generally larger size of NaI crystals, NaI:Tl detectors often have a higher detection efficiency for gamma rays than do HPGe detectors. Historically, the efficiency of a Ge detector was commonly compared to that of 3-in. diameter  3-in. long (3  3) right circular cylinder of NaI:Tl for 60Co gamma-ray energies (Tsoulfanidis and Landsberger, 2015). This standard continues today, where the efficiency of an HPGe detector is quoted in terms of a 3  3 NaI:Tl detector. For example, a 60% efficient HPGe detector has 60% of the efficiency that a 3  3 NaI:Tl detector has for 60Co gamma rays. HPGe detectors are much more expensive than NaI:Tl detectors; hence, they are best used when gamma-ray energy resolution is of importance whereas NaI:Tl detectors provide higher efficiency for measuring somewhat weak gamma-ray radiation fields.

Si detectors The problem with Li redistribution in Ge(Li) crystals when the detector warms does not apply to Si (Bertolini and Coche, 1968). Hence, Si(Li) detectors are still available in spite of their relatively poor efficiency for energetic photons. Because Si has a much lower average atomic number than Ge, the relative efficiency of Si(Li) to HPGe per unit thickness is significantly lower for electromagnetic radiation. However, for X-ray or gamma-ray energies less than about 30 keV, commercially available Si(Li) detectors are thick enough to provide performance comparable to HPGe detectors. For example, a 3–5 mm thick Si(Li) detector with a thin entrance window has an efficiency of almost 100% for 10-keV photons. Si(Li) detectors are preferred over HPGe detectors for low energy Xray measurements, primarily because of the lower energy X-ray escape peak features that appear in a Si(Li) detector spectrum as opposed to an HPGe detector spectrum (Goldstein et al., 1981). Further, background gamma rays, which complicate an X-ray spectrum, tend to interact more strongly in HPGe detectors than in Si(Li) detectors. Because a majority of the applications require a thin window, Si(Li) detectors are often manufactured with very thin beryllium windows. Typically, Si(Li) detectors are chilled with LN2 to reduce thermal leakage currents to allow optimum performance. High-purity Si detectors without Li drifting are also available, but are significantly smaller than HPGe and Si(Li) detectors. Such devices are typically only a few hundred microns thick and are designed for charged-particle spectroscopy. They range in diameter from one centimeter to several centimeters. The detectors are formed as diodes to reduce leakage currents, and use either a Schottky

Physical Sensors: Radiation Sensors

155

contact or a thin implanted pn junction contact to produce a rectifying diode. These devices are operated in reverse bias to reduce leakage currents. To reduce energy loss through the contact dead layer, the contacts and junctions are kept thin, typically being only a few hundred nanometers. Measurements are typically conducted in a vacuum chamber to reduce particle energy loss in air. Because the detectors are thin, they have low thermal generation of charge carriers and, consequently, can be operated at roomtemperature.

Alternative semiconductor detectors The fact that HPGe and Si(Li) detectors must be chilled during operation is a considerable inconvenience. Much research has been devoted to finding semiconductors that can be used at room temperature (Owens, 2012). The main requirement is that the bandgap energy be greater than 1.4 eV to reduce thermally generated leakage currents (McGregor and Hermon, 1997). For gamma-ray detection, the material is best composed of high atomic number elements. Consequently, there are only a few possibilities, all of which are compound semiconductors, meaning that they are composed of two or more elements. Three important semiconductors used for room-temperature operation are HgI2, CdTe, and CdZnTe. Mercuric iodide (HgI2) has been studied since the early 1970s as a candidate gamma-ray spectrometer material and has been used for commercial X-ray spectrometry analysis (Willig, 1971; Malm, 1972). The high atomic numbers of Hg (Z ¼ 80) and iodine (Z ¼ 53) make it attractive as an efficient gamma-ray absorber, and its large band gap of 2.13 eV allows it to be used as a room temperature gamma-ray spectrometer. Cadmium telluride (CdTe) has been studied since the mid-1960s as a candidate gamma-ray spectrometer material (Autagawa et al., 1967; Mayer, 1967). It has relatively good gamma-ray absorption efficiency with Cd (Z ¼ 48) and Te (Z ¼ 52). The band gap of 1.52 eV also allows CdTe to be operated at room temperature. There are commercial vendors of CdTe detectors but the devices are relatively small, typically a few mm thick with an area of a few mm2. CdTe detectors have been used for room-temperature-operated low-energy gamma-ray spectroscopy systems, and also for electronic personnel dosimeters. Cadmium zinc telluride (CdZnTe or CZT) has been studied as a gamma-ray spectrometer since 1992 (Butler et al., 1992; Doty et al., 1992). One of the most studied versions of CZT has 10% Zn, 40% Cd, and 50% Te molar concentrations, a composition which yields a material with a band-gap energy of approximately 1.6 eV. CZT detectors offer an excellent option for low-energy X-ray spectroscopy when cooling is not possible. Although the detectors are quite small compared to HPGe and Si(Li) detectors, they are manufactured in sizes ranging from 0.1 cm3 to 2.5 cm3. Because of their small size they perform best at gamma-ray energies below 1.0 MeV. Various clever electrode designs have been incorporated into new CZT detectors to improve their energy resolution, and CZT has become the most used compound semiconductor for gamma-ray spectroscopy (Owens, 2012; McGregor and Shultis, 2020).

Personnel dosimeters An important application of radiation detection is to measure the radiation dose received by workers in a radiation environment. Here the most common personnel dosimeters are discussed. These dosimeters can, or course, also be used as area monitors to measure the accumulated exposures at various locations in a radiation environment.

Photographic film Photographic film is one of the oldest radiation detection devices, having been used by Röntgen after his discovery of X-rays in 1895. The film badge consists of a packet of photographic film sealed in a holder with various attenuating filters. Ionizing radiation darkens the film as in the production of an X-ray image. The filtration is designed to modify the degree of film darkening, as nearly as possible, to a known function of gamma-ray exposure, independent of the energy of the incident gamma rays. After the badge is carried by a radiation worker for a period of time, the film is processed, along with calibration films with the same emulsion batch exposed to known radiation doses. The worker’s radiation dose for the period is assessed and ordinarily maintained in a lifetime record of exposure. In some cases, special attenuation filters are used to relate the darkened portions of the film to beta-particle or even neutron dose. Special badge holders in the form of rings or bracelets are used to monitor the radiation exposure of hands, wrists, and ankles (Cember, 1983).

Pocket ion chambers A compact form of ion chamber closely related to an electrometer is the pocket ion chamber. These devices are routinely used to measure the radiation dose received by the wearer (Cember, 1983). The device consists of a tube approximately 10 cm long and 1 cm in diameter. Inside the tube are two small metal-coated quartz fibers, each approximately 4 mm in diameter. One of the quartz fibers is stationary and the other is hinged, and both are inside an air cavity of the tube. The tube also has viewing optics such that when held up to a light, the observer can see the shadow of the hinged quartz fiber against a dose scale. The chamber is inserted into a power supply and charged so the two fibers, having like charges, are spread apart. This charged device can now be worn as a dosimeter. Electrons excited by radiation interactions in the air cavity are attracted to the quartz fibers and reduce the charge on the quartz

156

Physical Sensors: Radiation Sensors

fibers causing these fibers to move closer together (Eichholz and Poston, 1979). The change in location of the hinged fiber with respect to the metered display yields the dose (the scale maximum is usually 200 mR) (Cember, 1983).

TLDs and OSLs Thermoluminescent dosimeters (TLDs) and optically stimulated luminescent (OSL) dosimeters are used for radiation dosimetry measurements. Both TLD and OSL dosimeters are reusable and can be worn as personnel dosimeters. They can be read out for dose, cleared, and used again repeatedly. Select areas of an OSL dosimeter can be read out, with the unread areas still available for sampling at a later date. However, once a TLD dosimeter is read, the dose information is cleared and cannot be reevaluated. Dosimeters are checked periodically, between a few weeks to a month, and hence do not yield immediate information about radiation exposure. Rather, they are used to record the radiation dose accumulated by the wearer. TLDs and OSLs consist of small polycrystals held together by a binder. These polycrystals behave similar to a common scintillator with some important differences. Electrons elevated in energy from ionizing radiation can either recombine with holes or they can drop into traps. These traps are engineered into the bandgap by the introduction of specific impurities, and they can retain electrons for long periods of time. When a TLD is heated, the trapped electrons are released and can fall through luminescent centers to release light (McKeever, 1985). TLDs are gradually heated up to a maximum temperature ( 400  C) that ensures the complete release of all trapped electrons (Cameron et al., 1968). The light output is measured with a PMT as a function of the TLD temperature. The resulting displayed output (light vs temperature) is referred to as a glow curve (Cameron et al., 1968). Glow curves are unique to the TLD material type and impurity. The total light output varies linearly with the absorbed dose. If a TLD is not read out within a reasonable amount of time after the radiation exposure ( 1 or 2 months), a substantial number of electrons can return to the ground state. This effect is referred to as fading and causes some information about the radiation dose to be lost (Cameron et al., 1968). Although there are many useful TLD materials, the most common are lithium fluoride (LiF), calcium fluoride (CaF2), and calcium sulphate (CaSO4). LiF fluoride has a density similar to human tissue and is more popular than other TLD materials for personnel dosimetry. Neutron sensitive LiF TLDs are fabricated with enriched 6Li labeled TLD-600. Upon absorbing a neutron, energetic reaction products are released that are recorded by the TLD. Neutron insensitive TLDs can be made from LiF enriched with 7Li labeled TLD-700. Paired together, they can provide information on the gamma-ray and neutron doses. OSL dosimeters have become popular over the last few decades and have in many cases replaced TLDs (Bøtter-Jensen et al., 2003; Yukihara and McKeever, 2011). OSL dosimeters are usually carbon activated samples of Al2O3, and behave almost exactly the same as TLDs. Electrons are excited into the impurity traps where they stay lodged until the OSL dosimeter is read. However, instead of thermal excitation, trapped electrons in OSL dosimeters are dislodged with laser stimulation which can be directed to specific locations on the sample. The total light released by the OSL dosimeter is linearly related to the total accumulated dose. OSL dosimeters are more sensitive than TLDs to low levels of radiation and are capable of recording doses below 1 mrad.

Alternative detectors There are many novel radiation detectors, some unique, that are built to reveal the paths of radiation particles, to detect radiation through new physical mechanisms, or to record exotic radiations such as neutrinos and mesons. The description of the many specialized detectors, particularly those used in high-energy physics research, would require a separate chapter. Here just a few of these unusual detectors are introduced. Detailed discussions can be found in the literature and associated references (McGregor and Shultis, 2020).

Cloud chambers, bubble chambers, and superheated drop detectors Charles Wilson invented the cloud chamber in 1911, a device used to detect ionizing charged particles (Wilson, 1911, 1912). A cloud chamber is a sealed transparent container holding a supercooled, supersaturated vapor of water, alcohol, or aromatic hydrocarbon. Alpha particles or beta particles ionize the saturated vapor as they pass through it. The ionized vapor molecules form nucleation centers for condensation and a visible mist forms along the ionization track. Wilson shared the 1927 Nobel Prize in Physics for inventing the cloud chamber. Donald Glaser invented the bubble chamber in 1952, a type of detector closely related to the cloud chamber (Glaser, 1952). The bubble chamber is a vessel filled with a superheated transparent liquid used to detect charged particles as they move through the liquid. Charged particles deposit sufficient energy in the liquid so that it begins to boil along the ionization track, thereby forming a visible string of bubbles. Glaser received the 1960 Nobel Prize in Physics for inventing the bubble chamber. In 1979, Robert Apfel invented the superheated drop detector, which is similar to the cloud chamber and the bubble chamber (Apfel, 1979). A fluid can be superheated to temperatures and pressures corresponding to the vapor region in the phase diagram. This metastable state is fragile and typically short-lived due to the high number of microscopic particles or gas pockets normally present at the container surfaces. However, a liquid may be kept in steady-state superheated conditions by fractionating it into droplets and dispersing them in an immiscible host fluid, for example, suspending freon droplets in a gel solution. This procedure creates perfectly smooth spherical interfaces, free of nucleating impurities or irregularities. Radiation particles, including neutrons,

Physical Sensors: Radiation Sensors

157

interacting in the fluid can disrupt this balance between the liquid and vapor phase of a superheated droplet, causing it to burst into a sizable bubble of vapor. Bubble formation is measured from the volume of vapor expelled or by detecting individual vaporizations acoustically, i.e., “listening” to the radiation interactions.

Cryogenic detectors Radiation detectors that are becoming increasingly more interesting are the cryogenic or low-temperature devices (Enss, 2005). One such device is the microcalorimeter. These cryogenic devices sense the change in temperature of an absorber material caused by a single radiation quantum event in it. The device operates by attaching a small sample of thermally conductive material to a cryogenic refrigerator platform, and the sample is cooled to milli-Kelvin temperature levels with a refrigerator (such as a liquid-helium refrigerator or portable mechanical refrigerator). When a single quantum of radiation, such as a gamma ray or neutron, interacts in the chilled material, the temperature change DT in the material is measured and recorded. Because the temperature change is a linear function of the energy absorbed, a spectrum is accumulated as a function of temperature change. Extremely high resolution spectroscopy is possible, nearly 20 times better than achievable with Si(Li) detectors. Because the device must rapidly disperse the heat, the detector samples must be small. Consequently, cryogenic detectors work best for low-energy gamma rays and X-rays.

Radiation counting No measurement can be exact. There always are errors and uncertainties. With radiation measurements there is also the stochastic nature of radiation sources that leads to further uncertainties. In this section, such concerns are addressed.

Types of measurement uncertainties The analysis of any type of experimental data always requires an assessment of the uncertainties associated with each measurement. Without such an uncertainty estimate, the data have very limited value. There are several types of uncertainties associated with any measurement. These include stochastic and sampling uncertainties and errors as well as systematic errors. For example, the decay of radioactive atoms occurs stochastically so that a measurement of the number of decays in a fixed time interval has an inherent stochastic uncertainty. Repeated measurements would give slightly different results. Systematic errors are introduced by some constant bias or error in the measuring system and are often very difficult to assess since they arise from biases unknown to the experimenter. Sampling errors arise from making measurements on a population different from the one desired. These biases too are hard to detect, let alone quantify. Engineers and scientists must always be aware of the difference between accuracy and precision, even though popular usage often blurs or ignores the important distinction. Precision refers to the degree of measurement quantification as determined, for example, by the number of significant figures. Accuracy is a measure of how closely the measured value is to the true (and usually unknown) value. A very precise measurement may also be very inaccurate.

Uncertainty assignment based upon counting statistics Data measured with an ionizing-radiation detection system embody both random uncertainties and systematic errors. Uncertainty assignment requires knowledge of both. Even with a perfect measurement system capable of operating over a long time period without introducing significant systematic error, one must always consider the random nature of the data caused by the stochastic nature of the radiation source such as the decay of radionuclides in a sample. This section deals only with one component of the total uncertaintydthe random or statistical uncertainty. The binomial distribution describes the basic statistical distribution for the stochastic process of radioactive decay. However, the binomial probability distribution function is numerically difficult to use for a large number of radionuclides. For a large number of radionuclides, the probability of observing a single disintegration within a specified time interval is small (1). Also, the total expected number of observed disintegrations within that time is large (typically > 20). Hence, the Gaussian distribution approximates well the probability of observing x disintegrations about dx in some specified time period, namely, " # 1 ðx  mÞ2 dx; (1) GðxÞdx ¼ pffiffiffiffiffiffi exp  2s2 2p s where G(x) is the Gaussian or normal probability distribution function, m is the expected value of the random variable x being observed, and s2 is the variance of the distribution. Note that the number of counts x is treated as a continuous variable that can even assume unrealistic negative values, whereas the observed number of counts must always be a nonnegative integer. A typical Gaussian distribution is shown in Fig. 12, where it has been normalized to unit height. Data are routinely reported as x  sx where x is the average number of measured counts during a specified time period and sx is the standard deviation of x (Bevington and Robinson, 1992). When an error is reported as one standard deviation (or “one sigma”) it is called the standard error. For a single measurement x of a radioactive sample, the mean is estimated as x and it can be shown that

158

Physical Sensors: Radiation Sensors

Fig. 12 The Gaussian or normal distribution describes the expected probability distribution for radiation counting, where m is the expected number of counts observed and s is the standard deviation. From Radiation Detection: Concepts, Methods and Devices by D.S. McGregor and J.K. Shultis. Reproduced by permission of Taylor and Francis Group, LLC, a division of Informa PLC.

pffiffiffi the standard deviation, for x greater than about 20, is s ¼ x: Thus, for a single measurement, the reported value is x  s. If a Gaussian distribution can approximate the data, then s has a specific meaning, namely, there is a 68.3% probability that the next observed measurement will have a value that falls within the range x  s. Table 2 shows the relationship between the number of radiation induced counts recorded and the percent standard deviation. For example, a standard error of 1% or less would require recording at least 10,000 counts. Frequently, more than one measurement is recorded. If it is assumed the radiation source has constant activity, estimates of the mean and standard deviation can be obtained directly from the N measured values, x1, x2, ., xN. The experimental mean is the average value of the counts x ¼

N 1 X xi : N i¼1

(2)

Because si 2 ¼ xi, the total variance for a series of measurements is s2SUM ¼

N X

xi ¼ xN:

(3)

i¼1

For a series of N like measurements, assuming that the activity of the radioactive source remains fairly constant during the measurement series, it can be shown with Eqs. (2) and (3) that the standard deviation of x is qffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffi rffiffiffiffi s2SUM x xN sx ¼ : (4) ¼ ¼ N N N If another series of N measurements were to be taken, a new average value for x would be found, yet there would be a 68.3% chance that the new average would fall within the interval ranging from x  sx and x þ sx : The result of N measurements is reported as rffiffiffiffi x x  sx ¼ x  : (5) N Uncertainties are not always reported as one standard deviation. Sometimes a larger uncertainty is reported in order to increase the probability that the mean is included within the range between x  ks and x þ ks. As shown in Table 3, if the error is reported as

Table 2

Standard deviation of count data measured with a radiation detector operating in pulse mode.

Observed Total Counts

% Standard Deviation (1s error)

100 400 1100 2500 10,000

10% 5% 3% 2% 1%

Physical Sensors: Radiation Sensors Table 3

159

Probability interval relative to the number of standard deviations based on the Gaussian distribution.

Number of Standard Deviations (ks)

Probability Event is Observed within ks

0.67s 1.00s 1.65s 1.96s 3.00s

0.500 0.683 0.900 0.950 0.997

one standard deviation, the probability is 68.3%. However for an error range of x  1.65s to x þ 1.65s the probability increases to 90%. This would be reported as the 90% confidence interval. However, the convention is to report errors as one standard deviation. The Gaussian gamma-ray peak shape arises from the statistical fluctuations in the number of information (charge) carriers excited per monoenergetic gamma-ray interaction event. Typically the resolution of the gamma-ray peak from the detector is quoted as a function of the peak’s full width at half the maximum value (FWHM), which for a Gaussian distribution is 2.355s, where s is in units of energy. In terms of percent, the energy resolution of the spectrometer is calculated by dividing 235.5s by the gamma-ray energy.

Dead time All radiation detection systems operating in pulse mode have a limit on the maximum rate at which data can be recorded. The limiting component may be either the time response of the radiation detector such as that for a GM counter, or may be the resolving capability of the electronics. The true counting rate (n) for a detector with zero dead time losses is related to the recorded counting rate (m) for a detector with significant dead time losses by n¼

m ; 1  ms

(6)

where s is the dead time of the detector. Note that the term ms is the fraction of the time that the detector is unable to respond to additional ionization in the active volume of the detector. When designing an experiment, it is advisable to keep these losses to a minimum. If possible, this means that ms should be less than 0.05. For example, for a GM counter with a typical dead time of s ¼ 100 ms, the maximum count rate would be 500 counts/s.

Summary Since the discovery of X radiation by Rontgen in 1895, radiation detectors have undergone a variety of changes and improvements. These improvements include the changes from the initial use of scintillation phosphors such as BaPt(CN)4 to present-day bright scintillation spectrometers, the first Geiger counters described in 1908 to present-day multi-wire proportional counters, the first semiconductor conduction counters as described in 1941 and 1945 to present-day single-carrier compound semiconductor designs and arrays. Usually progress with radiation detector development was slow with long periods of small incremental improvements. At times a single important innovation would bring about a grand improvement in radiation detector development, for example, the photomultiplier tube that was commercially introduced in 1941 is a fundamental component of modern scintillation spectrometers. Such incremental development of radiation sensors punctuated by the introduction of novel new technology can be expected to continue. The vibrant surge in the development of radiation detectors and associated detector materials seen over the past 20 years resulted largely from new government programs and sponsorships such as, in the U.S., the 2002 Homeland Security Act, which had an avalanche effect on other governmental agencies to also focus on radiation sensors. As these programs now wane, radiation sensor development will still be driven by new applications in basic research (such as high energy and space physics), specialized commercial applications (such as medical imaging), military applications, and, indeed, by scientific curiosity. The universe is awash with radiation of different types and new radiation sensors will be needed for us to understand better the greater environment in which we exist.

References Apfel, R.E., 1979. The superheated drop detector. Nuclear Instruments & Methods 162, 603–608. Austin, L., Starke, H., 1902. Ueber die Reflexion der Kathodenstrahlen und eine damit verbundene neue Erscheinung secundärer Emission. Annalen der Physik 314, 271–292. Autagawa, W., Zanio, K., Mayer, J.W., 1967. CdTe as a gamma detector. Nuclear Instruments & Methods 55, 383–385. Becquerel, A.H., 1900. Deviation du Rayonnement du Radium dans un Champ Électrique. Compte Rendus 130, 809–815. Bertolini, G., Coche, A., 1968. Semiconductor Detectors. Wiley, New York.

160

Physical Sensors: Radiation Sensors

Bevington, P.R., Robinson, D.K., 1992. Data Reduction and Error Analysis for the Physical Sciences, 2nd edn. McGraw-Hill, New York. Birks, J.B., 1964. The Theory and Practice of Scintillation Counting. Pergamon Press, Oxford. Bøtter-Jensen, L., McKeever, S.W.S., Winkel, A.G., 2003. Optically Stimulated Luminescence Dosimetry. Elsevier, Amsterdam. Brown, W.L., Higinbotham, A.H., Miller, G.L. (Eds.), 1969. Semiconductor Nuclear-Particle Detectors and Circuits, Pub. 1593. National Academy of Sciences, Washington. Butler, J.F., Lingren, C.L., Doty, F.P., 1992. Cd1 – xZnxTe gamma ray detectors. IEEE Transactions on Nuclear Science 39, 605–609. Cameron, J.R., Suntharalingam, N., Kenney, G.N., 1968. Thermoluminescent Dosimetry. University of Wisconsin Press, Madison. Cember, H., 1983. Introduction to Health Physics, 2nd edn. Pergamon Press, New York. Chadwick, J., 1932. Existence of a neutron. Proceedings of the Royal Society A 136, 692–708. Cova, S., Lacaita, A., Ghioni, M., Ripamonti, G., Louis, T.A., 1989. 20-ps timing resolution with single-photon avalanche diodes. The Review of Scientific Instruments 60, 1104–1110. Dearnaley, G., Northrop, D.C., 1966. Semiconductor Counters for Nuclear Radiations, 2nd edn. Wiley, New York. Donati, S., 2000. Photodetectors. Prentice Hall, Upper Saddle River. Doty, F.P., Lingren, C.L., Butler, J.F., 1992. Cdi1 – xZnxTe gamma-ray detectors. IEEE Transactions on Nuclear Science NS-39, 605–609. Eichholz, G.G., Poston, J.W., 1979. Principles of Radiation Detection. Ann Arbor Science, Ann Arbor. Engstrom, R.W., et al., 1980. Photomultiplier Handbook. Burle Technologies, Inc, Lancaster. Cryogenic particle detectors. In: Enss, C. (Ed.), 2005. Topics in Applied Physics, vol. 99. Springer, New York. Evans, R.D., 1955. The Atomic Nucleus. McGraw-Hill, New York. Geiger, H., Müller, W., 1928. Das Elektronenzählrohr. Physikalishce Zeitschrift 29, 839–841. Glaser, D.A., 1952. Some effects of ionizing radiation on the formation of bubbles in liquids. Physics Review 87, 665. Goldstein, J.I., Newbury, D.E., Echlin, P., Joy, D.C., Fiori, C., Lifshin, E., 1981. Scanning Electron Microscopy and X-Ray Microanalysis. Plenum, New York. Grosshoeg, G., 1979. Neutron ionization chambers. Nuclear Instruments & Methods 162, 125–160. Haitz, R.H., 1961. Model of the electrical behavior of a microplasma. Journal of Applied Physics 35, 1370–1376. Hofstadter, R., 1949. The detection of gamma-rays with thallium-activated sodium iodide crystals. Physics Review 75, 796–810. Horrocks, D.L., Peng, C.T., 1971. Organic Scintillators and Liquid Scintillation Counting. Academic Press, New York. Kaplan, I., 1962. Nuclear Physics. Addison-Wesley, Reading. Korff, S.A., 1946. Electron and Nuclear Counters. Van Nostrand, New York. Lecoq, P., Annenkov, A., Gektin, A., Korzhik, M., Pedrini, C., 2006. Inorganic Scintillators for Detector Systems. Springer, Berlin. Liebson, S., 1947. The discharge mechanism of self-quenching Geiger-Mueller counters. Physics Review 72, 602–608. Liebson, S., Friedman, H., 1948. Self-quenching halogen-filled counters. The Review of Scientific Instruments 19, 303–306. Malm, H.L., 1972. A mercuric iodide gamma-ray spectrometer. IEEE Transactions on Nuclear Science NS-19, 293–295. Mayer, J.W., 1967. Evaluation of CdTe by nuclear particle measurements. Journal of Applied Physics 38, 296–301. McGregor, D.S., 2018. Materials for gamma-ray spectrometers: Inorganic scintillators. Annual Review of Materials Research 48, 13.1–13.33. McGregor, D.S., Hermon, H., 1997. Room temperature compound semiconductor radiation detectors. Nuclear Instruments & Methods A395, 101–124. McGregor, D.S., Shultis, J.K., 2020. Radiation Detection: Concepts, Methods, and Devices. CRC Press, Boca Raton. McIntyre, R.J., 1961. Theory of microplasma instability of silicon. Journal of Applied Physics 32, 983–995. McKeever, S.W.S., 1985. Thermoluminescence of Solids. Cambridge University Press, Cambridge. NCRP, 1985. A Handbook of Radioactivity Measurements Procedures, 2nd edn. National Council on Radiation Protection and Measurements, Bethesda, MD. NCRP Report No. 58. Owens, A., 2012. Compound Semiconductor Radiation Detectors. CRC Press, Boca Raton. Owens, A., 2019. Semiconductor Radiation Detectors. CRC Press, Boca Raton. Pell, E.M., 1960. Ion drift in an n-p junction. Journal of Applied Physics 31, 291–302. Price, W.J., 1964. Nuclear Radiation Detection, 2nd. edn. McGraw-Hill, New York. Renker, D., 2006. Geiger-mode avalanche photodiodes, history, properties and problems. Nuclear Instruments & Methods A567, 48–56. Ridley, B.K., 1982. Quantum Processes in Semiconductors. Clarendon Press, Oxford. Rodnyi, P.A., 1997. Physical Processes in Inorganic Scintillators. CRC Press, Boca Raton. Röntgen W (1895) Ueber eine neue Art von Strahlen. Vorläufige Mitteilung, Aus den Sitzungsberichten der Würzburger Physik.-medic Gesellschaft Wurzburg, 137–147. Röntgen W (1896) Eine neue Art von Strahlen. 2. Mitteilung, Aus den Sitzungsberichten der Würzburger Physik.-medic. Gesellschaft Würzburg, 11–17. Röntgen, W.C., Joffé, A., 1913. Über die Elektrizitätsleitung in einigen Kristallen und über den Einfluß der Bestrahlung darauf. Annalen der Physik 41, 449–498. Rutherford, E., 1899. VIII. Uranium radiation and the electrical conduction produced by it. Philosophical Magazine 47 (284), 109–163. Rutherford, E., Geiger, H., 1908. An electrical method of counting the number of a-particles from radio-active substances. Proceedings of the Royal Society A 81, 141–161. Sauli, F., 2014. Gaseous Radiation Detectors. Cambridge University Press, Cambridge. Shafroth, S.M. (Ed.), 1967. Scintillation Spectroscopy of Gamma Radiation. Gordon and Breach, London. Shultis, J.K., Faw, R.E., 2000. Radiation Shielding. American Nuclear Society, La Grange Park. Sitar, B., Merson, G.I., Chechin, V.A., Budagov, Y.A., 1993. Ionization Measurements in High Energy Physics. Springer, Berlin. Slepian J (1923) U.S. Patent 1,450,265. Stetter, G., 1941. Durch Korpuskularstrahlen in Kristallen Hervorgerufene Elektronenleitung. Verhandlungen der Deutschen Physikalischen Gesellschaft 22, 13–14. Sze, S.M., 1981. Physics of Semiconductor Devices, 2nd edn. Wiley, New York. Thomson, J.J., 1897. Cathode rays. Philosophical Magazine 44, 293–316. Tsoulfanidis, N., Landsberger, S., 2015. Measurement and Detection of Radiation, 4th edn. CRC Press, Boca Raton. Van Heerden, P.J., 1945. The Crystal Counter, A New Instrument in Nuclear Physics. Noorde-Hollandsche Uitgevers-Mij, Amsterdam. Villard, P., 1900a. Sur la Réflexion et la Réfraction des Rayons Cathodiques et des Rayons Déviables du Radium. Comptes Rendus 130, 1010–1012. Villard, P., 1900b. Sur le Rayonnement du Radium. Comptes Rendus 130, 1178–1179. von Broser, I., Kallman, H., 1947a. Uber die Anregung von Leuchtstoffen Durch Schnelle Korpuskularteilchen I. Zeitschrift für Naturforschung 2a, 439–440. von Broser, I., Kallman, H., 1947b. Über den Elementarprozess der Lichtanregung in Leuchtstoffen Durch Alpha-Teilchen Schnelle Elektronen und Gamma-Quanten II. Zeitschrift für Naturforschung 2a, 642–650. Wilkinson, D.H., 1950. Ionization Chambers and Counters. Cambridge University Press, Cambridge. Willig, W.R., 1971. Mercury iodide as a gamma spectrometer. Nuclear Instruments & Methods 96, 615–616. Wilson, C.T.R., 1911. On a method of making visible the paths of ionising particles through a gas. Proceedings of the Royal Society of London A 85, 285–288. Wilson, C.T.R., 1912. On an expansion apparatus for making visible the tracks of ionising particles in gases and some results obtained by its use. Proceedings of the Royal Society of London A 87, 277–292. Yen, W.M., Shionoya, S., Yamamoto, H., 2007. Fundamentals of Phosphors. CRC Press, Boca Raton. Yukihara, E.G., McKeever, S.W.S., 2011. Optically Stimulated Luminescence Fundamentals and Applications. Wiley, New York.

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors Abdelghani Ghanama,b, Hasna Mohammadia, Aziz Aminea, Naoufel Haddourb, and Franc¸ois Buretb, a Equipe Analyses Chimiques et Biocapteurs, Laboratoire Génie des Procédés et Environnement, Faculté des Sciences et Techniques, Université Hassan II de Casablanca, Mohammedia, Morocco; and b Laboratoire Ampère, École Centrale de Lyon, CNRS UMR 5005, Université de Lyon, Écully, France © 2023 Elsevier Ltd. All rights reserved.

General introduction Voltammetric techniques Principle of voltammetry Theory of voltammetry Types of voltammetry Linear sweep voltammetry/cyclic voltammetry Pulse and square wave voltammetry Selection of voltammetric technique Voltammetric sensors and biosensors Indicator (or working) electrode StationarydTransient mode Direct or indirect (bio)sensing Recent applications of voltammetric sensors and biosensors Amperometric techniques Amperometry principles and characteristics Amperometric sensors Amperometric sensors applications Gas sensors H2O2 sensors Hydrazine sensors Nitrite sensors Phenol sensors Clinical component sensors Amperometric biosensors applications Enzyme-based sensors Immunosensors DNA biosensors Conclusions References

161 162 162 163 163 163 166 166 167 167 168 169 169 170 171 171 172 172 172 172 172 172 173 173 174 175 175 175 176

Abstract Sensors based on voltammetry and amperometry are involved in several fields such as clinical, food, environmental, pharmaceutical, and industrial sectors. The present article describes the principle of the voltammetry and amperometry and highlights the practical applications of various analytical techniques such as cyclic voltammetry, differential pulse voltammetry, square wave voltammetry and chronoamperometry. Thanks to the modification of the sensors with nanomaterials, polymers, mediators, and bioreceptors sensitivity and selectivity are much improved. Biosensors have demonstrated their excellent ability to be integrated easily into batch or flow injection modes, wearable or miniaturized electrochemical devices, allowing them to be adapted to specific applications.

General introduction Increasing worldwide concern about environmental pollution, food contamination, drug safety, and security issues has resulted in a strong and urgent need to develop rapid, sensitive, and reliable methods for monitoring and detecting dangerous chemical compounds. Hence, a variety of conventional and sophisticated physico-chemical methods are currently available in the market for this purpose, including atomic absorption spectroscopy (AAS), inductively coupled plasma mass spectrometry (ICP-MS), HPLC coupled mass spectrometry (HPLC-MS), and UV-visible spectrophotometry. These techniques are undoubtedly reliable and accurate, but unfortunately are time-consuming and they require heavy, expensive, and highly sophisticated equipment and sample preparation steps. Moreover, due to their lack of sensitivity and selectivity,

Encyclopedia of Sensors and Biosensors, Volume 1

https://doi.org/10.1016/B978-0-12-822548-6.00032-7

161

162

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

they often require prior sample preparation such as purification, pre-concentration of the target analyte, etc. (Baltussen et al., 2002). Considerable effort has therefore been devoted to developing simple, real-time monitoring and early warning analytical methods to quickly identify potentially dangerous compounds (molecules, ions) in our environment. In the first section of this article, we focused on voltammetric techniques (VT): based on scanned potential. Voltammetric sensors for environmental, food, and clinical applications using the most commonly used VT are also covered. Moreover, a brief and detailed description of the VT, i.e., cyclic voltammetry (CV), linear sweep voltammetry (LSV), differential pulse voltammetry (DPV), and square-wave voltammetry (SWV) is presented, citing the theoretical aspects and principles as well as the measurement characteristics, advantages, and limitations of each technique. This section also critically discusses a range of recent advances in electroanalysis and bio-electroanalysis using the VT. In the second section, the amperometric technique will be discussed.

Voltammetric techniques The VT are a group of electroanalytical techniques working at controlled variable potential (i s 0) in which the current is measured as a function of applied potential. The first VT was developed from the discovery of polarography in 1922 by the Czech chemist Jaroslav Heyrovsky, who was awarded the Nobel Prize in Chemistry in 1959. Further to this discovery, during the 1960s and 1970s, theories, methods, and instrumentation were developed in the voltammetry field, increasing sensitivity and extending the repertoire of electroanalytical methods (Bockris and Reddy, 2012). Indeed, some of the most developed VT are shown in Fig. 1. Unlike static potentiometric techniques, voltammetry imposes an external energy source on the solution to be analyzed in order to provoke/force an electrochemical reaction (Faradic reaction). Indeed, for any type of analyte, as long as it can be reduced or oxidized electrochemically, the resulting faradic current (iF) is proportional to its concentration. The advantages of VT include an excellent selectivity and sensitivity with a broad concentration range towards several analytes; a lower limit of detection (LOD); a wide range of temperatures; fast analysis times (few seconds); simultaneous detection; speciation; and the ability to determine kinetic parameters and estimate the mechanisms of chemical and/or electrochemical reactions.

Principle of voltammetry The instrumentation (Fig. 2) for carrying out voltammetry typically involves a three-electrode system (working, reference, and counter/auxiliary electrodes) known as an electrochemical cell (Bard and Faulkner, 2001a). Besides, the common characteristic of all VT is the application of a variable potential between a reference electrode (RE), e.g., Ag/AgCl, and an indicator electrode frequently called a working electrode (WE), where the oxidation-reduction reaction (Ox þ ne– / Red) takes place. In practice, to ensure that the generated current does not flow through the RE and its impedance is increased, a third electrode named counter electrode (CE) or auxiliary electrode (e.g., generally based on stainless steel, noble metal, or carbon) is required. Employing

Fig. 1 Classification tree for some electrochemical techniques. The specific techniques are shown in blue, the experimental conditions are shown in green, and the analytical signals are shown in red.

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

163

Fig. 2 General setup for voltammetric measurements and schematic representation of a typical three-electrode electrochemical cell setup used in voltammetric techniques, where an inert gas (N2, Ar, etc.) is bubbled into the electrolyte solution containing the analyte of interest.

a potentiostat, this new electronic three-electrode system helps maintain a constant and controlled potential at the WE with respect to the RE. The principle then is based on the measurement of anodic/cathodic current provided by the oxidation/reduction of an electroactive species at a selected potential window. In brief, by scanning (increasing/decreasing) the potential at a well-defined range where the reduction or oxidation of analyte could be realized, a resulting current as a function of potential curve is obtained, i ¼ f(E), the resulting plot called voltammogram. Additionally, as the recorded voltammogram reflects the electrochemical behaviour of redox species involved in the oxidation or reduction reaction, it provides quantitative/qualitative information about it (Maloy, 1983).

Theory of voltammetry Each three-electrode system has its thermodynamic characteristics, particularly the electrode potential, an important reference point for the system because it defines its equilibrium state where no current (i ¼ 0) flows in the cell. Indeed, at zero current, the potential window over which the WE can work without causing electrolysis of the electrolyte (electroactivity window limits) is also essential. The larger it will be, and the more attractive the electrode used in electroanalysis will be. In the presence of the redox couple (analyte), the thermodynamic potential (Eth) of WE depends on its standard potential (E ) as well as the ratio of the concentrations of the oxidized and reduced forms of the couple at the electrode surface. The thermodynamic potential is described by the Nernst equation (1).   RT C Eth ¼ E  ln red (1) nF Cox where R is the gas constant (8.3144 J mol–1 K–1), T is the temperature (K), n is the number of electrons transferred, F is Faraday constant (96480 C mol–1), E is the standard potential of the redox couple, Cred/Cox is the ratio of redox species (Ox and Red form) at the electrode surface. Therefore, any additional potential (beyond the thermodynamic requirements) will trigger an electrochemical reaction at a specific rate. The rate may depend on the electrode material, the electrolyte, and the electroactive behaviour of the analyte. The difference between the thermodynamic potential of a WE, immersed in an electrolyte solution containing a redox species (depolarizers), and that characterizes the new equilibrium state when a iF flows in the cell, is called polarization. The extent of polarization is measured by the overpotential, h (Eq. 2). Current-potential curves, especially those obtained under steady-state conditions, are sometimes referred to as polarization curves (Brett and Oliveira Brett, 1993; Bard and Faulkner, 2001b). h ¼ E  Eth

(2)

where h (V) is the overpotential measured at zero current; a kinetic parameter determining the reversibility of the redox couple, E (V) is the applied potential necessary to trigger a cathodic or anodic electrochemical reaction, compared to Eth (V) (thermodynamic potential at zero current) depending on the standard potential E of redox species, their concentrations, and the electrolytic medium.

Types of voltammetry Linear sweep voltammetry/cyclic voltammetry LSV and CV are the most commonly used potential sweep techniques. The Potential-excitation signal of the WE (Fig. 3a) is based on a linear potential wave-form in the case of LSV and triangular for CV, where for both techniques, the potential is changed linearly with time (t). The potential change rate is known as the scan rate v (dE/dt), which typically varies between 1 mV s–1 and 1 V s–1 (Brett

164

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

Fig. 3 Details for LSV and CV. (a) time-dependent potential excitation signal showing the applied initial potential (Ei) and the final/switching potential (Ef) for the forward (blue line) and the reverse scan (dotted line), ti and tf are initial and final time, respectively. (b) corresponding I ¼f(E) curve (LSV, only one scan; CV, both direct and reverse scan). (c) A typical cyclic voltammogram of current as a function of applied potential. (d) change in concentration profile as a function of time of oxidized and reduced species at electrode surface for near peak potential. Scan rate, dE/ dt. Adapted from Bard AJ and Faulkner LR (2001a) Basic potential step methods. Electrochemical Methods: Fundamentals and Applications, 2.

and Brett, 1998). In addition, once the oxidizing/reducing potential of the species in solution is reached, the iF rises to the maximum. Thereafter, without stirring the solution, the current decreases due to a depletion of the species that can be oxidized/ reduced near the electrode surface (Fig. 3d), revealing an oxidation/reduction peak (Fig. 3b). For LSV, the potential explored is scanned from the initial potential (Ei) to the final potential (Ef), either towards more positive potentials or more negative potentials. However, in CV a completed scan in both directions by reversing the potential at the end of the direct scan (first scan), which in turn becomes the switching/starting potential of the reverse scan. In comparison to LSV, CV has become a very popular technique for electrochemical studies of new electrochemically active systems. CV provides qualitative and quantitative information about the analyte’s electrochemical behaviour under study (ion, molecule, etc.). In addition, CV is the most powerful and versatile VT to indicate: in a one-step the oxidation/reduction process, a number of redox reactions associated chemical reactions and their mechanistic study (Hawley et al., 1967), thermodynamic and kinetic parameters such as the number of electron changes (n), the heterogeneous rate constant (k ), Gibb’s free energy (G), and diffusion coefficient (D) (Brett and Oliveira Brett, 1993). Furthermore, to quantify any chemical substance (ion, molecule) for the first time, analytical electrochemists widely use CV as a preliminary technique for initial characterization. Fig. 3c depicts a typical cyclic voltammogram (I vs. E) for a reversible process. The characteristic parameters in a cyclic voltammogram are as follows:

• ip, a: peak current of anodic oxidation, • Ep, a: peak potential of anodic oxidation, • ip, c: peak current of cathodic reduction, • Ep, c: peak potential of cathodic reduction; • O Ep ¼ r Ep, a  Ep, c r: Peak-to-peak separation. In addition, three electron transfer processes can be distinguished by CV: the reversible (Nernstian), the irreversible, and the quasireversible process. The O Epvalues are almost 60 mV expected for a mono-electronic reversible process, whereas they are greater than 60 mV for a quasi-reversible process.

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

165

Electrochemical active surface area determination (A) For an electrochemical system controlled only by diffusion (fast electron transfer), the electrochemically active surface area (A) of a WE was estimated by changing the scan rate based on the Randles-Sevcik equation (Randles, 1948) in a redox probe solution (i.e., [Fe(CN)6/Fe(CN)6]3 – /4–). The determination of A of the sensor is crucial to evaluate its electroactivity before and after its modification prior to being used in voltammetric or amperometric measurements. Adsorption or diffusion-controlled process By changing the scan rate (i.e., from 5 to 500 mV s–1) using CV measurement in an electrolyte solution containing the electroactive species of interest (such as ferri/ferrocyanide), it might be possible to identify whether the electron transfer process is controlled by diffusion or by adsorption (i.e., diffusion if I ¼ f(v1/2) presented better linearity than that of I ¼ f(v)), corresponding to the adsorption process (Sharp et al., 1979). Sometimes, it is not easy to distinguish them as I ¼ f(v) and I ¼ f(v1/2) graphs exhibited similar linearity. Therefore, to identify the most dominant process, Eq. (3) is used to provide accurate information. If the slope (a) obtained is almost unity, the electrochemical reaction is governed by adsorption. Additionally, knowing the number of electrons n and A, the surface coverage concentration G of the adsorbed/desorbed electroactive species can be accurately calculated from the slope of Eq. (4).   2 2   n F GA (3) log Ip ¼ a logðvÞ þ log 4RT Ip ¼

n2 F2 GA v 4RT

(4)

where G represents the surface coverage concentration (mol cm–2).

For soluble electroactive species and diffusion-controlled processes, Nicholson (1965) has published, for n O Ep  200 mV, a mathematical treatment, which made it possible to use CV to evaluate the electrode kinetics. Therefore, measuring the heterogeneous standard rate constant (k ) of electron transfer (cm s–1) as well as estimate of reversibility limit in order to distinguish the

reversible from the irreversible redox process. In this case, in CV, the O Ep increases as a function of dimensionless kinetic parameter J, which equals: rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  a=2 rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi D RT RT ¼ k J ¼ k O (5) ¼ Ck  v1=2 pnFDO v pnFDv DR



ð  0:6288 þ 0:0021 XÞ where X ¼ nOEp ð1  0:017 XÞ

(6)

Where the short form of Eq. (5) is by assuming symmetrical redox kinetics (a z 0.5), mono-electronic process, and the diffusion coefficient of oxidized (DO) and reduced form (DR) to be equal. Briefly, k is calculated from the slope of J ¼ f (v 1/2). Firstly, n O Ep values are calculated from the CV by changing the scan v. Secondly, as J depends on

O Ep, it can be determined from

Eq. (6) based on the data, J ¼ f (n O Ep),provided by Nicholson’s theory. On the contrary, for adsorbed electroactive species and adsorption-controlled processes, Laviron (Laviron, 1979) published the mathematical treatment allowing the use of CV to determine the rate constant of electron transfer (kET) and the transfer coefficient (a) of adsorbed electroactive species (Eqs. 7–9). Indeed, such adsorption-controlled process yields an ideal “Nernstian behavior” of surface-confined species that is reflected by a symmetrical cyclic voltammogram (O Ep z 0) with a peak half-width (O Ep, 1/2)   ¼ 90:6 independent of 3:53 RT nF n . In practice, this ideal behaviour is approached for a relatively slow scan rate (or v tends to

0), and for an adsorbed layer that shows no intermolecular interactions and fast electron transfers. However, when v increases,

O Ep increases and the electron transfer process tends towards the “quasi-reversible process”. For adsorbed electroactive species and adsorption-controlled processes showing n O Ep > 200

mV, Laviron proposed the

following equations: Ep;c ¼ E þ Sc logðvÞ where Sc ¼  Ep;a ¼ E þ Sa logðvÞ where Sa ¼  logk ET ¼ alogð1  aÞ þ ð1Þ  aÞ loga  log

2:3 RT anF

2:3 RT ð1  aÞnF

 RT nF  að1  aÞ OEp nFv 2:3RT

(7)

(8)

(9)

166

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors Sa a Sc ¼ 1a. Therefore, based on the 90:6), the k is obtained from Eq. (9). ET n

where a can be easily obtained from the slope ratio where calculated from the peak half-width when v tends to 0 (

calculated value of a and the n

Chemical reaction mechanisms One of the most important cyclic voltammetry applications is for a qualitative investigation of some electron transfer-coupled chemical reactions and their mechanism. Generally, when an electroactive species can undergo one or more chemical reactions preceding or following the electron transfer, it significantly changes the peak current responses and their ratio

Ip;r Ip;f

, as well as the

shape of the cyclic voltammogram. Generally, two letters, “E” and “C,” have been used to classify the reaction mechanisms of “electrochemical” and “chemical” steps, respectively. The most frequently studied and encountered situations:

• • • •

EC mechanism: electron transfer followed by chemical reaction, CE mechanism: chemical reaction preceded by electron transfer, ECE mechanism: Electrochemical-chemical-electrochemical, ECC mechanism: Electrochemical-chemical-chemical.

A chemical or biological reaction could generate such an electroactive species (e.g., the use of enzymes), its electrochemical reaction could be useful to quantify the concentration of the starting substrate, which is, in many cases, non-electroactive.

Pulse and square wave voltammetry CV and LSV are excellent for diagnosing redox species, but the severe limitation is their low sensitivity. Their LOD is limited only from 10–3 to around 10–5 M (Wang, 2006), making them unsuitable to be incorporated with voltammetric sensors. Hence, is a result of the relative contributions of iF and capacitive (or charging) currents (iC) to the total current (iT) of the electrochemical cell. Discrimination between iF and iC, thus increasing voltammetric sensors’ sensitivity, is achievable with pulse methods (Brett and Oliveira Brett, 1993). The most commonly used are DPV and SWV, which are much better candidates than CV and LSV selected as techniques of voltammetric measurements. The pulse techniques are all based on chronoamperometry. It is important to note that 1 , Cottrell’s law, more details in section Amperometry principles and charin chronoamperometry the iC, decreases faster than iF (in t1=2 acteristics). Moreover, when a potential pulse is applied to an electrode, both iC and iF decrease over time with rates essentially different (i e., mainly, iF decreases faster than iF). Further rejection of the iC is achieved with DPV, limiting the size of the applied pulse amplitude (Fig. 4DPVa). The current in DPV is sampled twice at each pulse period, at the beginning (1) and the end (2) of the pulse. Subtracting the current value in (1) from that in (2) allows the background current to be rejected, leaving only the iF. As shown in Fig. 4DPVb, the difference between the two current values is recorded and displayed as a function of the applied potential. Better iC rejection means that low detection

limits and a high signal-to-noise ratio may be easily reached. The main advantage of the DPV is that the pulse potential (O Ep) can separate the responses of compounds that oxidize or reduce to close potentials and subsequently improve the selectivity and allows simultaneous detection. SWV is one of the fastest pulse VT (can reach 1.0 V s–1) because there is no renewal of the diffusion layer, unlike DPV. The excitation signal in SWV consists of a symmetrical square-wave pulse having constant amplitude Esw, superimposed to a staircase waveform, as shown in Fig. 4SWVc. The net current (DI [I1–I2) in SWV comes by subtracting the measured current (2) at the end of the forward pulse from the measured current (1) at the end of the reverse pulse. Because of the wide amplitude of the square wave, for a reversible reduction, the reduced electroactive species formed at the electrode during the forward pulse is re-oxidized by the reverse pulse (Fig. 4SWVd). Therefore, the sensitivity of this method is increased compared to DPV (Dogan-Topal et al., 2010). Furthermore, typical SWV measurements take only 1–5 s, whereas DPV requires much longer analysis times at about 2–4 min (Mirceski et al., 2007). The optimum value of the square wave amplitude Esw is 50 n mV, irrespective of the values of DE and or s. Decreasing this value causes a loss of sensitivity, with no diminution of the peak width; increasing Esw causes peak broadening, with no enhancement in sensitivity. The other advantage of SWV is that The DI is, in such cases, larger than either forward or reverse currents (Fig. 4SWV d and e), so the height of the peak is usually quite easy to read, thus increasing the accuracy. Moreover, for a reversible redox couple, the theoretical I-E curve (Fig. 4SWVd) takes a symmetrical peak shape. So, due to the presence of reverse current (I2), the resulting current DI is much higher than the forward current (I1) whereas, as shown in Fig. 4SWVe, for an irreversible system, the reverse current is low, indicating that DI is not much improved compared to the forward current (I1) (Brett and Oliveira Brett, 1993). The effect of irreversibility is to shift the peak potential to more negative values (for reduction), decrease the peak height, and enlarge it. These effects are common to all voltammetric methods but remain negligible for SWV than for DPV.

Selection of voltammetric technique The choice of VT depends mainly on the characteristics of the sample under study, in particular, the analyte type (ion or molecule), the expected concentration of the analyte and its electrochemical behaviour (such as the reversibility, adsorption, or diffusioncontrolled process), and the desired accuracy and precision.

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

167

Fig. 4 Illustration of the pulse techniques (DPV and SWV) and their usual values of the different parameters, showing (A) and (C) the potentialexcitation signal as a function of time as well as the current sampling for DPV and SWV, respectively, (B) the corresponding typical I vs. E for DPV plot. In SWV, (D) and (E) represent theoretical I–E responses of a reversible and an irreversible system, respectively. s is the pulse period, tp is the pulse time, Estep is a potential step, OEp is the pulse amplitude, ESW is the square wave amplitude, and OI is the resulting current. Adapted from Osteryoung JG and Osteryoung RA (1985) Square wave voltammetry. Analytical Chemistry 57(1): 101–110; Brett C and Oliveira Brett AM (1993) Electrochemistry: Principles, Methods, and Applications (544.6 BRE).

As mentioned, DPV and SWV are better suited for quantitative determination due to their potential to discriminate iC. Pulse techniques frequently are designed for quantitative analysis at a very low concentration, while CV and LSV for electrochemical diagnosis and characterization of electrochemical sensors. In consequence, the choice of VT to use depends on various factors, such as the concentration range, the electrode materials, the accuracy and precision necessary, the required response time etc.

Voltammetric sensors and biosensors The history of electrochemical sensors, including biosensors, is a well-known subgroup, which began in 1960 following the first work carried out by Clark and Lyon to detect glucose (Clark and Lyons, 1962). These studies were deepened around 1967 following the birth of the first biosensor made by Updicke and Hicks (Updike and Hicks, 1967) and was dedicated to glucose determination. According to the IUPAC definition, a biosensor is an integrated bioreceptor-transducer device that combines biological sensing materials (Nucleic acid, enzyme, etc.,) with a transducer. The interaction between the analyte and the bioreceptor is then converted into a measurable electrical signal (displayed or playback) through the transducer. In contrast to a biosensor, an electrochemical sensor is a device that transforms the electrochemical interaction between the analyte and abiotic receptor over the electrode surface into an analytically useful signal. The abiotic receptor (i.e., Metallic nanoparticles, Carbon nanomaterials, conducting polymers, nanocomposites, etc.) is the fundamental component of a sensor in the sensing which monitors and detects molecular changes, and the transducer, which transforms the changes observed by the receptor into measurable signals. Furthermore, if the electrochemical technique applied to sensors’ working is voltammetry (i.e., DPV, SWV, etc.), the electrochemical sensors are called voltammetric sensors.

Indicator (or working) electrode Electrochemical reaction (oxidation or reduction) takes place on the working (or indicator) electrode (WE), which will then lead to the detection of the target analyte. In addition to the VT used, supporting electrolyte, and pH, the WE material is considered one of the most critical parameters affecting the iF. Hence, WE should be crucially well selected to widen the potential window for several analytes a and improve the sensitivity, specificity, reliability, and accuracy of the detection.

168

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

On the other hand, all WEs are limited by their oxidation in the electrolyte medium at a certain level towards positive potentials. The depolarizers can also be oxidized or reduced electrochemically when WE’s potential reaches a certain value. Basically, all electroor bioelectro-analysts work with aqueous solutions; thus, the potential window (electroactivity region) of the WE is usually limited by the oxidation and reduction of water (Fig. 5). Beyond that, the water oxidizes (H2O / 1/2O2 þ 2Hþ þ 2e–) or reduces (H2O þ 2e– / H2 þ 2OH) in its turn. These two potential values, depending on the supporting electrolyte, the pH and the RE (Rouessac et al., 2019). In the sensing area, the WE (or so named bare electrodes) can be used without any further modification. However, lack of sensitivity and selectivity imposes a sample preparation process (i.e., solid-phase extraction column) or modified electrode with an appropriate modifying agent such as carbon nanomaterials (i.e., Carbon black), thiolated MNPs (Mandil et al., 2010), MIPs (Zamora-Gálvez et al., 2016), metallic NPs (Pd, Pt, Ag NPs) as catalyst (Laghrib et al., 2019; Wu et al., 2019), and nanocomposites (Oularbi et al., 2017).

StationarydTransient mode As known, all voltammetric techniques work with an unstirred solution, and the mass transport of electroactive species is governed only by diffusion. In potential scanning methods, it should always be kept in mind that the scan rate is a parameter to be taken into account when establishing the stationary regime: if the observation time of the electrochemical signal is too short (i.e., high scan rate), the stationary regime of the transport will not have time to be established, herein we talk about transient voltammetry. Two possibilities exist to obtain a stationary regime:

• •

Rotating disk electrode (RDE) Ultramicroelectrodes (UME)

Voltammetry measurements performed with a RDE are part of the hydrodynamic methods, for which the convection of the solution is forced and well controlled. The rotation of the electrode induces in the diffusion layer a perpetual renewal of the electroactive species. The limit current (Il), for a reversible system, is proportional to the square root of the angular velocity, as described by Levich’s equation (Bard and Faulkner, 2001a): Il ¼ 0:62nFAD2=3 u1=2 C v1=6

(10)

where Il (Ampere) is the limit current, u is the angular velocity ¼ 2pf (rads s–1), where f is the number of revolutions per second, v is the kinematic viscosity (cm2 s–1) UME appeared at the end of 1970s with the Wightman’s works (Dayton et al., 1980). The original idea was the miniaturization of electrodes to be implanted in vivo for the detection of biologically interesting species (Neurotransmitters such as Dopamine). UMEs are defined by the size of their surface area, less than 50 mm. They have well-known advantages, such as small capacitive currents, reduction of resistive effects in the cell, a stationary or quasi-stationary regime reaches very quickly without any need to impose convection on the electrochemical system (Bond et al., 1989; Huang et al., 2009). In a stationary regime, the voltammograms do not show peaks but platters.

Fig. 5 Potential window of Boron-doped diamond electrode at pH 2.0 and platinum, mercury, and carbon electrodes in various supporting electrolytes.

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

169

Direct or indirect (bio)sensing The detection of molecules or biomolecules using voltammetric sensors can be performed directly or indirectly. Indeed, many electroactive molecules such as sulfonamide (Lahcen et al., 2016) give an electrochemical response whose peak current is directly proportional to their concentration, while for some heavy metals such as Cr(VI), its detection can be made indirectly through electrochemical oxidation of diphenylcarbazide (DPC) (Hilali et al., 2018). In brief, a nonelectroactive complex, DPC-Cr, is formed once the Cr (VI) is injected into the solution containing DPC, thus decreasing the initial response of the latter. The difference between responses before and after the injection of Cr(VI), is proportional to the concentration of Cr(VI). A number of biomolecules can be electrochemically oxidized or reduced for direct detection, such as oxygen and most neurotransmitters (DA, EP). However, most cancer biomarkers (microRNAs) require indirect detection. The presence of a target molecule generates electroactive species for detection via electrochemical sensors or the capture of target alters the ability of the electrode to facilitate electron transfer process (on/off detection). Furthermore, for indirect detection, specific recognition elements (DNA, Enzyme, bacteria, etc.) are required for each target of interest, which is often obtained by immobilization/modification of electrode surfaces. Biosensors based on enzyme inhibition also offer indirect detection of the analyte (inhibitor) (El Harrad et al., 2018; Islam and Channon, 2020).

Recent applications of voltammetric sensors and biosensors Over the last decades, voltammetric sensors have developed considerably. The results obtained with these small detection devices have shown their enormous potential compared to other conventional analytical techniques in terms of short analysis time, excellent sensitivity with a wide range of concentrations, very LOD, and improved selectivity and speciation of particular metallic ions with two different oxidation states (i.e., Cr(III) and Cr(VI)). Such a voltammetric sensor, before being incorporated in wearable devices, commercialized and applied in the food, environmental, clinical, pharmaceutical, industrial sector, etc., must first fulfill several important requirements (Fig. 6). Voltammetric sensors and biosensors can be applied in various areas, including Point-of-Care (POC) diagnosis, early ultrasensitive diagnosis of diseases (such as cancers, diabetes), drug monitoring, and rapid and accurate detection of environmental and food pollutants. In addition to fouling (Ghanam et al., 2017), Bisphenol A (BPA) has a relatively high overpotential preventing its direct electrooxidation due to increased interference, causing low selectivity of the sensor. However, one of the solutions of electrode fouling was recently proposed by Rojas et al. (2020). Briefly, the authors described a new sensor based on transition metal dichalcogenide (MX2) nanosheet (where M: Mo or W; X: S or Se) towards electrochemical oxidation of catechol-containing flavonoids. Moreover, the selected nanomaterial has successfully shown its excellent potential as a promising solution by solving sensor fouling problems compared to carbon-based electrodes. Furthermore, the authors have clarified the mechanism behind the antifouling resistance of this new generation of nanomaterial. Since the first commercialized glucose biosensor in 1973 by Yellow Springs Instrument, enormous efforts have been made in the development of biosensors academically, bringing some of them to be commercialized in the market, drug stores, and Website. Moreover, Bahadır and Sezgintürk (2015) have reported a review (110 references) summarizing the widely used and commercially available biosensors. Indeed, biosensors have shown their excellent capability to be integrated easily with low cost, high sensitivity, fast response, and easy use for on-site and POC biosensing. The global biosensor market size was estimated to reach US$19.6 billion in 2019. Many chemical and/or biochemical analytes in clinical (cancer biomarkers playing a crucial role in early disease diagnosis, glucose for diabetes diagnosis, lactate, ethanol, cholesterol, creatinine), environmental (heavy metals, pesticides, biochemical oxygen demand, nitrate), food (i.e., glutamate, glutamine, sucrose, lactose, alcohol, ascorbic acid), and biothreat/biowarfare (i.e., Bacillus anthracis, Salmonella, Botulinum toxin) sector can be measured successfully by using commercial biosensors (Bahadır and Sezgintürk, 2015). Glucometer and i-STAT are one of the most popular models used for POC biosensing as well as for

Fig. 6

Illustration of important requirements and features of an excellent (bio)sensor.

170

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

monitoring of food (fruits, vegetables) maturity. For further examples or in-depth discussion, we recommend the review (Bahadır and Sezgintürk, 2015). The three generations of electrochemical glucose biosensors are all based on glucose oxidation into gluconic acid in the presence of oxygen by means of the enzyme glucose oxidase (GOx) immobilized on the electrode surface. Unfortunately, the high cost, the poor stability of GOx and their complicated immobilization procedure and chemical deformation during their manufacturing process, and the storage remain the significant limitations facing electrochemical glucose biosensors (Ren et al., 2019). However, the introduction of non-biological catalysts has attracted much attention for glucose detection at physiological conditions, as, in principle, to overcome first to third generation glucose biosensors’ limitations. This new approach has given rise to a 4th generation based on non-enzymatic detection with low cost, high stability, and biocompatibility. More recently, in a review (131 references), Adeel et al. (2020) presented recent advances in biocompatible nonenzymatic electrochemical sensors based on metallic or their alloy nanocatalysts for the direct electrocatalytic oxidation of glucose. The review also discussed the wearability of some nonenzymatic glucose sensors for POC testing and on-site monitoring of glucose. Cancer has become a worrying and prevalent cause of illness and death in recent years. It is now the second leading cause of death worldwide, costing millions of lives every year. Electrochemical biosensors are considered a powerful tool for ultrasensitive detection of cancer biomarkers that play a crucial role in early disease diagnosis (Yammouri et al., 2019). Cancer biomarkers include a variety of biomolecular entities from cellular and subcellular sources such as nucleic acids (e.g., DNA and RNA), functional subgroups of proteins (e.g., enzymes, glycoproteins, tumor antigens, and receptors), circulating tumor cells (CTCs) as well as cells with overexpressed tumor antigens and extracellular vesicles (e.g., tumor-derived exosomes) (Islam and Channon, 2020). Biosensors for nucleic acid biomarkers are suitable for analysis either in centralized laboratories or at the POC. Thanks to their high sensitivity, many biosensors can detect endogenous nucleic acid concentrations with no amplification steps, thus reducing analysis time and simplifying the analysis process (Islam and Channon, 2020). Determination is mainly based on the hybridization of target sequences (such as biomarker) to complementary receptor probes (mostly DNA probes) on the electrode surface. Moreover, based on the intrinsic electrochemical properties of the nucleic bases, the hybridization process could be easily measured by voltammetry (DPV, CV) using a redox probe/electroactive indicator such as hexaammineruthenium (III) chloride ([Ru(NH3)6]3þ) and ferri/ferrocyanide ([Fe(CN)6]3  /4 ) to obtain an electrochemical signal. The miniaturization of such an electrochemical device for on-site (bio)analysis is limited by its components such as potentiostat, electrodes, sample volume, etc. The constant progress in printing technology has led to the development of miniaturized platforms for (bio)sensing in the field of voltammetric sensors and biosensors. In addition, screen-printed electrode (SPE), as an example, were first introduced in the 1990s, and their reliability, mass production, reproducibility, suitability, and low cost have contributed to fulfilling this requirement. In a recent review (155 references), some of us described a new emerging electrode based on laserderived graphene (LDG) technology (Lahcen et al., 2020). The electrode produced is mainly called laser-scribed graphene (LSG) or laser-induced graphene (LIG). Thanks to their unique properties (such as mask-free production, mechanical stability, porosity, 3D graphene structure, high electrical conductivity with fast electron mobility, and large surface area), the new generation LSG showed their potential to be a good candidate for sensors and biosensors application and to be integrated into wearable electrochemical devices for real-time monitoring and point-of-care testing. In turn, Sinha et al. (2020) presented a chapter (39 references) on MXene-based sensors and biosensors as next-generation detection platforms. MXene has emerged as a unique 2D material that mainly includes early transition metal carbides, nitrides, and carbonitrides. They are produced by etching out A layer from a 3D structure consisting of MAX (Mn þ 1AXn), where M is an early d-transition metal, A is the main group sp element, and X is C or N. Moreover, this new material has demonstrated promising features for sensing and biosensing applications towards several including biomolecules (DA, AA, H2O2, etc.), environmental contaminants (Pb2þ, Cd2þ, etc.), and gaseous molecules (NO2, CO2, O2, etc.). Furthermore, MXen-modified electrodes have proven to be effective transducers for immobilizing various biological receptors (e.g., enzymes) on its surface. In a mini-review (106 references), Zhao et al. (2019) described the recent advances in metal-organic frameworks (MOF)-based electrochemical sensors and biosensors. Thanks to their unique properties, including chemical stability, tunable structure and property, and ultrahigh porosity, MOFs have demonstrated their excellent features for sensing and biosensing applications.

Amperometric techniques The present section discusses amperometry as an analytical technique. This technique examines a range of recent applications based on different amperometric transducers. Amperometry is an electroanalytical technique measuring the current resulting from the electrochemical oxidation or reduction of an electroactive species at a constant potential. The resulting steady-state current (iF) is proportional to the bulk of the analyte concentration. A three-electrode system constitutes the most used configuration for this kind of measurements. The WE, is maintained under the optimal constant potential versus RE. The WE can be metal (frequently used Pt, Au) or carbon-based material (carbon paste electrode, glassy carbon, graphite, carbon fiber, or more recently, the screen-printed). The benefits of controlled-potential techniques are represented in high sensitivity and selectivity towards electroactive species and a wide linear concentration range. Various electrodes permit analyses in unusual environments and can be integrated into portable and low-cost instrumentation.

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

171

Amperometry principles and characteristics Under amperometric measurements, the iF generated from the reaction is proportional to the analyte’s concentration under the optimal constant potential (Wang, 2006). Therefore, the electron transfer occurs at a diffusion-controlled rate, so the process becomes controlled by mass transfer (Kissinger and Heineman, 2018). The diffusion-controlled current (i) depends on electron transfer number n, the surface area A of the WE, analyte concentration C, and the Faraday number F (96480 C mol–1), by the following equation (Janata, 2009): i ¼ nFAKC

(11)

where C is a concentration surrounding the surface of the electrode and K constant, which includes diffusion coefficient and diffusion thickness layer. This technique, therefore, represents a system with three electrodes with solution stirring. The passage of a iF occurs during an exchange of electrons by oxidation or reduction of species in solution on the surface of the WE and of the CE. These reactions occur at a characteristic minimum potential for each electrochemical species and thus make it possible to quantify them in solution. When the potential of the WE is stepped, and the resulting current from faradaic processes occurring at the electrode is checked with time (a very short time), this leads to the development of a variant of amperometric technique, known as chronoamperometry. In this case, the transport of the reactants and products to or from electrode surface is governed by diffusion. Indeed, the solution should be unstirred in chronoamperometric measurements. Meanwhile, the mass transport under these conditions is only controlled by diffusion. The current as a function of the time reflects the gradient of the concentration t in the vicinity of the electrode surface. As a result, the current decreases with time, as given by the Cottrell equation: i ðt Þ ¼

nFACD1=2 ¼ kt1=2 p1=2 t1=2

(12)

where n, F, A, C, D, and t are the number of electrons, Faraday’s constant, the surface area, the concentration, the diffusion coefficient, and time, respectively. Over the years, and thanks to increasing interest in electrochemical research, these concepts have been further extended to develop a more advanced variant known as double-potential-step chronoamperometry. Accordingly, amperometry is of considerable interest in electrochemical sensor and biosensors researches (Amine and Mohammadi, 2019). The electrochemical amperometry offers the capacity to distinguish selectively between many electroactive species in solution by judicious selection of the applied potential and electrode material choice. Moreover, the potential must be carefully chosen to avoid interference problems. In amperometry, the potential is not scanned, accordingly does not lead to any voltammogram. The adequate potential value is usually selected using data from cyclic voltammetry, depending on the desired transformation (oxidation or reduction) from the region where the analyte of interest gives a limiting current plateau (Wang, 2006). More details are given in the voltammetric section.

Amperometric sensors Oxygen electrodes are among the most employed amperometric sensors (Qlark Jr, 1956). Their applications cover biology, medicine, industry, energy production, and safety (Janata, 2009). Likewise, the conditions under which the oxygen sensors have to operate range from high temperature (400–1600  C) sensing in gas phase to liquid media sensing at room temperature. A major advance in the amperometric oxygen sensors performance is achieved by placing both the cathode and the anode behind the oxygenpermeable membrane. This sensor is recognized as the Clark oxygen electrode. Several membrane materials were used, namely Teflon, polyethylene, and silicon rubber among others. Good selectivity is obtained by choosing the material of this membrane depending on the conditions of the application. The amperometric technique using unmodified bare electrodes be regarded as an interesting method, especially when high sensitivity is not required. This approach is appreciated by using a simpler system, resulting in reduced costs for both production and use and long-term stability. However, using a bare electrode has some disadvantages, such as over-potential, fouling, and interferences problems. To overcome the aforementioned problems, different strategies using modified electrodes have been integrated. Recently, a wide variety of mediators, polymers, and nanomaterials, notably carbon-based nanomaterials, metallic nanoparticles, oxide nanoparticles, semiconductor nanoparticles, and even composite nanomaterials, are applied in amperometric sensing (George et al., 2018; Sivakumar et al., 2019). Since nanomaterials have exceptional catalytic properties, they improved the electron transfer. Because of their important surface area and a high level of free energy, nanomaterials can strongly adsorbing biomolecules and therefore have an important function in immobilizing biomolecules used for biosensor design. Several combinations of nanomaterials have been studied for improving the analytical performances of the amperometric sensors (Sivakumar et al., 2019). The amperometric transducers have been coupled with several systems such as the flow injection analysis to reduce the fouling problem, allow simultaneous detection, and increase the throughput as required in routine analysis. More recently, advances in electrochemical sensing and wireless communication technologies have triggered the conception of wireless (bio)sensors with applications in the fields of health, defense, sport, environment, and agriculture. Amperometry is clearly one of the dominant transduction methods (Kassal et al., 2018; Silvester, 2019).

172

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

Amperometric sensors applications Several applications based on the chemical modified sensors use emerged in recent years.

Gas sensors The amperometric gas sensors are well established for industrial safety. The use of amperometric electrochemical gas sensors for monitoring inorganic gases that affect urban air quality has been widely used since the 1970s (Chang et al., 1993). From this time, the use of amperometric electrochemical gas sensors for monitoring inorganic gases that affect urban air quality has been widely used. A fast responding amperometric CO2 microsensor with ionic liquid aprotic solvent electrolytes has been developed (Fapyane and Revsbech, 2020). The performance of a Clark-type CO2 sensor was significantly improved in the presence of dimethylformamide. A sensor optimized for high sensitivity exhibited a linear range within 0.6–1.7 mM with the LOD of 0.2 mM. Sensing behaviour of room temperature amperometric H2 sensor with Pd electrodeposited from ionic liquid electrolyte as modified electrode has been reported by Jayanthi et al. (2017). Linear range of response of the sensor to H2 was in the concentration range 1– 5%. The amperometric gas sensor with atomic Au–Polyaniline-Pt composite has been designed and studied by Chakraborty et al. (2020). The selectivity of atomic metal catalysts for various aroma compounds such as geraniol, nerol and their electro-oxidation to manufacture smell sensing arrays has been investigated. The solid polymer-bonded electrode for amperometric N2O detection at low concentrations has been developed by Persaud and Phillips (2000). The new electrode is adequate for nitrous oxide gas sensors with amplified sensitivity owing to direct contact of the gas with the electrode. An amperometric NH3 sensor based on Sm2Zr2O7 solid electrolyte and SrM2O4 sensing electrodes has been developed by Cong et al. (2020). This work shows that the developed sensor exhibits high sensitivity (trace detection, ppm), quicker response, excellent selectivity, and stability, demonstrating great application prospects in automotive equipment. An amperometric hydrogen sulfide gas (H2S) sensor based on a Pt-Ni alloy electrode and a proton-conducting membrane has been investigated by Li et al. (2020). the authors have obtained low LOD of 200 ppb.

H2O2 sensors Hydrogen peroxide (H2O2) is considered an important mediator in various application fields such as pharmaceutical analyses, food control, clinical diagnostic, industry production, and environmental monitoring. It plays an important role in various biological processes. In addition, H2O2 is a byproduct of some biochemical reactions catalyzed by oxidase enzymes such as GOx. Several amperometric strategies based on solid electrodes, especially platinum electrode, have been used for H2O2 determination. These strategies presented some drawbacks, such as the over-potential and interference problems, which can be solved using mediators and nanomaterials Consequently, electrocatalytic reduction of H2O2 at low applied potential was successfully achieved using nanozyme composed of nanomaterial and Prussian Blue modified electrodes (Komkova et al., 2020). In addition, the H2O2 reduction at around 0 V avoided interference from other electrochemical species. Alternatively, Emir et al. (2020) modified pencil graphite electrode with Cupric-Neocuproine Complex as a new redox mediator for the electrocatalytic oxidation of H2O2 in flow injection analysis. This new strategy enhanced electrocatalytic oxidation of H2O2 rather than its reduction and proved better overpotential elimination and low LOD (0.4 mM of H2O2). The nonenzymatic electrochemical H2O2 sensor based on molybdenum oxides has been reported (Li et al., 2019). The developed amperometric nano-sensor displays the broad linear detection from 0.92 mM to 2.46 mM and the excellent calculated sensitivity (391.39 mA mM–1 cm–2) with a low LOD of 0.31 mM. The sensor was applied for the determination of H2O2 in human serum samples.

Hydrazine sensors A sensitive and rapid flow injection amperometric hydrazine sensor using an electrodeposited gold nanoparticle on graphite pencil electrode has been reported (Teoman et al., 2019). This sensor displays a linear calibration curve between the flow injection amperometric current and hydrazine concentration within the concentration range from 0.01 to 100 mM with a LOD of 0.002 mM. The flow injection amperometric sensor has been successfully used for the determination of hydrazine in water samples.

Nitrite sensors Recently, a nitrite flow-injection amperometric sensor for gunshot residue screening has been developed (Promsuwan et al., 2020). A glassy carbon electrode modified with a composite of palladium particles and glassy carbon microspheres has been manufactured to detect gunshot residue via nitrite. The developed sensor showed a linear response to nitrite between 0.10 mmol L–1 and 4.0 mmol L–1, with a LOD and LOQ of 0.030 and 0.11 mmol L–1, respectively.

Phenol sensors A liquid chromatographic separation with pulsed amperometric analysis of phenolic acids at a glassy carbon electrode is reported (Freitas et al., 2018). Linear ranges of the analytes, in mg L–1, were of 0.018–18 (gallic acid), 0.146–19 (vanillic acid), 0.13–17

Chemical Sensors: Voltammetric and Amperometric Electrochemical Sensors

173

(caffeic acid), 0.016–16 (ferulic acid), and 0.008–17 (p-coumaric acid), respectively and LOD ranged from 1.6 to 97 mg L–1. Vanillic and caffeic acid concentrations as 19 and 7.8 mg L–1 were detected in sugarcane vinasse samples.

Clinical component sensors Simultaneous analysis of paracetamol, acetylsalicylic acid, and caffeine in pharmaceutical tablets, using boron-doped diamond coupled to batch injection analysis system with multiple pulse amperometric detection has been investigated by Machado Alencar et al. (2018). Under optimized conditions, the proposed method exhibits fast responses with suitable precision. The MoS2 nanosheets/reduced graphene oxide composite paper electrode showed high current for folic acid oxidation at less positive over-potential (from 950 to 730 mV) (Kırans¸an and Topçu, 2018). The TiO2/reduced graphene oxide nanocomposite as efficient ascorbic acid amperometric sensor has been studied (Harraz et al., 2019). The developed sensor exhibits good sensitivity of 1.061 mA mM– 1 cm–2, low LOD 1.19 mM, a linear range of concentration (25–725 mM), and a rapid response time < 5 s. In addition, such sensor showed good stability, repeatability, and high selectivity in the presence of interfering species such as dopamine, glucose, uric acid, and citric acid. The developed senor showed satisfactory detection result towards a commercially available vitamin C supplement. The electrochemical oxidation of glucose at copper-modified electrodes is reported by Mohammadi et al. (2004). After parameters optimization, amperometric determination of glucose was performed at þ 0.30 vs. Ag/AgCl in 0.15 M NaOH. Good linearity is obtained in the range of 4–1000 mM of glucose. Zhu et al. (2019) used another approach, based on ultrafine platinum nanoparticles grown on a 3D graphene framework via a hydrothermal method. This sensor works at low applied potential ( 0.15 V vs. Ag/AgCl) and allows linear responses in the 0.1 mM to 0.01 mM and 0.01–20 mM glucose concentration range, with 30 nM as LOD. It was applied to the rapid determination of glucose in human serum samples. A nanostructured MXene-based biomimetic enzymes for amperometric detection of superoxide anions from HepG2 cells has been described (Zheng et al., 2019). This approach facilitates the electron transfer between O2• and electrode. The sensor, best operated at þ 0.75 V (vs. Ag/AgCl), showed a rapid amperometric response with a LOD of 0.5 nM and an analytical range that extends from 2.5 nM to 14 mM. Typical amperometric sensors are resumed in Table 1, and their various analytical performances are highlighted.

Amperometric biosensors applications Several categories of biosensors may be distinguished, depending on the nature of the biological recognition process, such as biocatalytic biosensors, immunosensors, and another category that can now be added such as bio bio-receptor or biomimetic receptor.

Table 1

Typical amperometric sensors.

Analyte

Electrode

CO2

Ionic liquid aprotic solvent electrolytes/Ag Cathode

Detection limit (DL) linear range (LR) DL: 0.2 mM

Sample

Distinctive

Urea

0.72 V

Alcohols NPrOH/ Atomic Au–Polyaniline-Pt Concentration of 1600 ppm – iPrOH vapors composite of pure alcohol in air NH3 Sm2Zr2O7 solid electrolyte and LR: 25–500 ppm – SrM2O4 (M¼ Sm, La, Gd, Y) H2S H2O2 Hydrazine Nitrite Ascorbic acid Anions from HepG2

Pt-Ni alloy electrode/proton DL: 200 ppb conducting membrane pencil graphite electrode/ LR: 1000–10,000 mM Cupric-Neocuproine mediator DL: 0.4 mM porous carbon-NiO nanocomposites glassy carbon electrode/ palladium particles/glassy carbon microspheres TiO2/reduced graphene oxide nanocomposite MXene-based biomimetic enzyme

LR: 0.5 mM to 12 mM DL: 1.5 mM LR: 0.10 mM to 4 mM DL: 0.11 mM LR: 25–725 mM DL: 1.19 mM LR: 2.5 nM to 14 mM DL: 0.5 nM

– Cosmetic samples, hair oxidation cream/antiseptic solution Milk and disinfectant samples lake and tap waters Gunshot residue Vitamin C supplement Human serum

References

Fapyane and Revsbech (2020) Smell sensing arrays Chakraborty et al. (2020) Cong et al. Automotive (2020) equipment. At 600–800  C Detection at room Li et al. (2020) temperature Flow-injection Emir et al. (2020) amperometric sensing 0.0 V Sivakumar et al. (2019) Flow-injection Promsuwan amperometric et al. (2020) sensor Rapid response Harraz et al. time 1960s) semiconductor-based conductometric gas sensor devices (Neri, 2015), sometimes also referred to as chemiresistive sensors. In this kind of modern devices the physi/chemisorption process occurring at the semiconductor surface, mediated or not by a suitable catalyst, is directly converted into an electronic transfer between the target analyte and the semiconductor itself, whose electric conductivity is then observed to change as a result of the interaction. The kinetic of the process relies, in general, on three basic steps: (a) adsorption, (b), diffusion, and (c) desorption of the target analyte. Gas sensor device response time (tr), that is, the time for the device signal to reach its maximum as long as the target analyte concentration in the reference environment increases, mainly depends on steps (a) and (b). When the analyte concentration in the gas phase is reduced, the analyte molecules are desorbed from the sensing material, driven by the concentration gradient and a recovery time (tR) can be defined. Selectivity and sensitivity can be addressed, in this type of devices, by a suitable choice of the semiconductor and its fabrication process, of the catalyst and of the operating temperature. Modern conductometric gas sensor devices are mostly based on two main classes of chemiresistive materials: (i) semiconductor metal-oxides (MOX) and (ii) nonoxide materials (polymers, carbon nanotubes and 2D materials such as graphene and transition metal dichalcogenides). Due to their low cost, good portability and high sensitivity MOX based gas sensors, first commercialized in the 1960s in Japan with the names Taguchi (the inventor) or Figaro (the company’s name) (Taguchi, 1971) rapidly diffused in domestic and industrial fields for monitoring flammable as well as toxic gases and are nowadays of reference for any other type of gas sensor devices. Their widespread use is only limited by the quite high operating temperatures, which can also cause sensor response temporal drifts and device stability issues. Conversely, nonoxide based gas sensors, with their possible lower operating temperatures and, in turn, their low power absorption, represent a valid alternative. Gas sensors based on these last materials are however still far from a full industrial development and commercialization.

Sensing materials and operating mechanisms In the following sections, the physico-chemical operating mechanism of both semiconductor metal-oxides and nonoxide materials, such as polymers, carbon nanotubes and 2D materials are discussed.

Metal oxides Metal oxides are a class of inorganic, semiconducting materials that comprises a huge number of compounds, many of which have been extensively investigated in gas sensing applications and where a direct physical and/or chemical interaction between target analytes and the semiconductor enables changes of the electrical properties and gives rise to a sensing signal. These interactions can take place at the surface or in the bulk of the sensing layer and can range from physical interactions (adsorption, diffusion, swelling) to chemical reactions (hydrogen bonding, covalent bonding, catalysis). MOX can be fabricated by means of a number of different processing techniques such as: precipitation, hydrothermal/solvothermal, electrospinning, sputtering, sol-gel, thermal oxidation, thermal evaporation, physical vapor deposition (PVD), chemical vapor deposition (CVD), atomic layer deposition, reactive ion etching, molecular beam epitaxy. Their growth mechanisms depend on some conditions such as the phase of the precursor (vapor or solution), the presence, or not, of catalyst, and the template (Li et al., 2015).

192

Chemical Sensors: Conductometric Gas Sensors Table 1

band gap of metal oxides. MOX exhibit band gap values that span from those typical of insulators (e.g., MgO and Al2O3), to wideband gap (e.g., TiO2, SnO2) to narrow-band gap semiconductors (e.g., CuO) (Wang et al. (2010), Korotcenkov (2007)). Metal oxide

Eg (eV)

MgO, CaO, Al2O3, SiO2, TeO2 SrO, Y2O3, ZrO2 BaO, La2O3, CeO2, GaO3 TiO2, Ta2O5, ZnO, In2O3, SnO2 V2O5, Cr2O3, WO3, NiO, Fe2O3 Co3O4, PdO, CuO, Sb2O3

>6 5–6 4–5 3–4 2–3 1–2

In Table 1, some examples of metal oxides are listed. They exhibit band gap (Eg) values that span from those typical of insulators (e.g., MgO and Al2O3), to wide-band gap (e.g., TiO2, SnO2) and narrow-band gap semiconductors (e.g., CuO). In a MOX sensing material the chemical reactivity and the conductivity mechanisms are mainly determined by the material operating temperature, as it is schematically reported in Fig. 3. The temperature indeed affects the response of the sensor both through the chemical reaction rate between the detected species and the sensing surface, and through the diffusion rate of the gas molecule: in particular, the response of MOX-based sensors increases with operating temperature, reaching a maximum at a given Tmax. For temperatures above this value, the response decreases. This behavior may be explained based on the kinetics and mechanism of gas adsorption and desorption on metal oxide surface. The sensing response results from a balance of two competing phenomena. At T < Tmax, the molecules of the target gas do not possess sufficient thermal energy to react with the oxygen absorbed by the surface species, hence, in such temperature range, the chemical reaction rate between the sensing film and the target gas acts as a limiting factor on the detection response. Conversely, the higher temperatures favor the oxygen desorption from the metal oxide film, leading to a reduced surface coverage of oxygen. Hence, at T > Tmax bulk effects may dominate over the surface chemistry and consequently the sensing response is limited by the diffusion rate of gas molecules (Gadkari et al., 2010). There is therefore an optimal temperature at which both processes are present and, concomitantly, a maximization of the response is observed; such temperature is specific for each target gas and normally ranges between 250 C–350  C. At such a range of temperatures, when a MOX is heated in ambient air, oxygen molecules are adsorbed on its surface and get ionized, by attracting electrons from the conduction band of MOX as it is shown in Eq. (2) and schematically depicted in Fig. 4. O2 ð gasÞ þ e 4O 2 ðadsorbÞ

(2)

The oxygen ions on the surface of a MOX are extremely reactive toward target gas molecules and the electrons involved can move back from the surface reactive sites into the bulk conduction band of the MOX. This change of the carrier concentration modifies material conductivity and can be therefore recorded by the external measuring circuit in terms of a conductance change. Material conductivity variation depends on the MOX semiconductor type. In n-type MOX, the majority carriers are electrons that are captured by the adsorbed oxygen ions according to the mechanism depicted in Eq. (2). This causes a decrease in the material carrier concentration and thus an increase of the resistance of n-type MOX based sensors. If an n-type MOX sensor is exposed to

Gas

Metal oxide

Gas Metal oxide

OperatingTemperature°C

Gas Redox

Bulk diffusion

200°C

Metal oxide

Chemisorption

400°C

700°C

Fig. 3 Operating temperature and gas-MOX reaction. Adsorption/desorption and bulk diffusion mechanisms compete to control material conductivity according to the operating temperature.

Chemical Sensors: Conductometric Gas Sensors

193

Fig. 4 Activation of redox reaction: oxygen is adsorbed on MOX surface and gets ionized, by attracting electrons from the conduction band with the formation of a depletion region.

a reducing gas, (i.e., a gas that easily donates an electron to an oxidizing agent in a redox reaction), the electrons transferred by the chemical reaction are given back to the conduction band leading to a decrease of the sensor resistance as it is shown in Fig. 5. Viceversa, in a p-type MOX, the majority carriers are holes. In air, when the environmental oxygen adsorbed on the surface of p-type MOX attracts electrons from the valence band, new holes are generated. This process results in increasing the number of charge carriers, which leads to a decrease of the sensor device resistance. When the p-type MOX sensor reacts with a reducing gas, electrons are injected back into the valence band where they recombine with the holes, finally reducing the number of free carriers. As a result, the sensor device resistance is now observed to increase. An opposite behavior is observed in oxidizing environments. The gas sensing properties of differently prepared MOX toward polluting gases (NO2, NO, H2S, NH3, CO, CH4) are reported in the following Table 2 with data obtained from Chavali and Nikolova (2019) and Wetchakun et al. (2011). According to Eq. (2) and to the above depicted sensor operating mechanisms, device sensitivity is expected to strongly depend on the number of material sites that can react with the surrounding environment, that is, on the material surface structure and morphology. The sensitivity of a MOX based conductometric gas sensor can therefore be increased by changing the sensing material microstructure where the adsorption-diffusion-desorption process occurs. In this respect, most of the MOX fabrication technique for gas sensor applications are optimized to realize partially sintered grains, connected to each other by necks (Sun et al., 2012) as it is shown in Fig. 6, the interconnected grains forming larger aggregates connected to each other by grain boundaries. Generally, there

Fig. 5 Sensing response of n-type (left panel) and p-type MOX (right panel) when exposed to reducing gas. In n-type MOX, the majority carriers (i.e., electrons) are captured by the adsorbed oxygen ions, thus decreasing the carrier concentration and increasing the resistance. When n-type MOX sensor is exposed to a reducing gas, the electrons transferred by the chemical reaction are given back to the conduction band leading to a decrease of the sensor resistance. For the p-type MOX based sensor, the majority carriers are holes. The electrons extracted by the adsorbed oxygen increase the concentration of holes inside so that the resistance of p-type MOX decreases, while electrons released by reducing gases provoke a resistance increase. Image adapted from Wang Y (2016). Room Temperature Gas Sensing Using Pure and Modified Metal Oxide Nanowires. Thesis.

194

Chemical Sensors: Conductometric Gas Sensors Table 2

SnO2

ZnO

In2O3 WO3

CuO CdO

Examples of MOX gas sensing properties for different gases. Analyte

Detection range

Optimal temperature ( C)

H2 NO2 NH3 CO CH4 H2 H2 S NO2 NH3 CO H2 S NO2 CO H2 S NO2 NO NH3 NO2 CO NO2

10–100 ppm 0.5 ppm 100 ppb–50 ppm 100 ppb 1% 200–500 ppm 50 ppb–10 ppm 5–50 ppm 50 ppm 300 ppm 200 ppb–50 ppm 1 ppm 100 ppm 1 ppm 50 ppb–1 ppm 0 –440 ppm 10 ppb 2 ppm 1–30 ppm 1 ppm

150–300 200 RT–300 300 700 RT RT RT–300 RT 350 RT–300 RT–250 300 250 130–200 150–250 260 300 RT–300 100

Data from Chavali M S and Nikolova MP (2019) Metal oxide nanoparticles and their applications in nanotechnology. SN Applied Sciences 1(6): 607; Wetchakun K, Samerjai T, Tamaekong N, Liewhiran C, Siriwong C, et al. (2011) Semiconducting metal oxides as sensors for environmentally hazardous gases. Sensors and Actuators B: Chemical 160 (1): 580–591.

Fig. 6 Schematic model of the effect of the crystallite size on the sensitivity of metal oxide gas sensors. Image adapted from Sun YF, Liu SB, Meng FL, Liu JY, Jin Z, Kong LT, Liu JH (2012) Metal oxide nanostructures and their gas sensing properties: A review. Sensors 12: 2610–2631.

are three cases with respect to the relationships between the grain size (D) and the width of the space-charge layer (L), which are classified in terms of boundary control, neck control and grain control. For a large grain where D (grain size) > 2 L (thickness of the space charge layer), the conductivity is limited by Schottky barrier at grain boundaries (known as grain boundary control). If D ¼ 2 L, conductivity is limited by necks between grains (known as neck control) and if D < 2 L, conductivity is influenced by

Chemical Sensors: Conductometric Gas Sensors

195

each grain (known as grain control). In the latter situation, the electric properties of the material are dominated by the surface and by its reaction with the surrounding environment so that excellent response to the target gases can be expected. Yamazoe was one of the first to focus his research activity on this issue, observing that the sensitivity of sintered SnO2 elements, with particles having diameter in the range of 4–27 nm, strongly increased when the average size of the crystallites D decreased, as it is shown in Fig. 7 (Yamazoe, 1991). In this graph, the sensitivities are reported as Ra/Rg, where Ra is the resistance measured in the reference environment and Rg is the resistance measured in the target analyte. It can be clearly observed that for grain size D > 20 nm, the sensitivity is nearly independent on size, but below 20 nm it increases with decreasing grain size, an effect which becomes even more remarkable below 10 nm. Therefore, in general, more porous is a sensing material, more easily gas molecules can penetrate deep into the material itself, resulting in an increase of the material reactive sites available and, therefore, of the sensor response, that is the sensor operating mechanism becomes more and more efficient, the smaller are the grains of the material. Surface decoration with noble metal nanoparticles is another powerful method for enhancing gas sensing characteristics. Noble metal additives such as Pd, Pt, Au or Ag on an oxide surface or in its volume have been long investigated in this respect and their promoting effects in MOX gas sensors has been widely verified as it is reported in Table 3, with data adapted from Wetchakun et al. (2011). In Fig. 8, it is shown how small amounts of noble metals, such as Pd, Pt and Au, added to MOX, such as SnO2 and ZnO, increase the sensitivity according to a sensitization mechanisms that can be chemical or electronic in nature. The electronic sensitization, realized by covering n-type metal oxides surface with metal particles, results in an electron migration from metal oxide to metal, so decreasing the electrons concentration in the former and causing an enlargement of the space charge region. At elevated temperatures, the air oxygen reacts with the metal surface extracting electrons from it; the metal, in turn, extracts electrons from the metal oxide, leading to a further increase of the space-charge depth. In the case of flammable gases, such as hydrogen, the enhanced sensitivity is determined by the reaction with the oxygen adsorbates both on the metal and on the metal oxide surface. In this case, therefore, the promoting effect arises mainly from the change in the oxidation state of the loaded material. In the chemical sensitization, the promoting effect is due to the spill-over induced by metal particles, that facilitates the reaction of the inflammable gases with oxygen adsorbates. Besides, the presence of these catalytic additives dissociate at the surface oxygen molecules from the gas phase, so increasing the oxygen atoms amount able to migrate to the surface of the metal oxide through spill-over effect. In this way, the additive exerts a sort of remote control on the catalytic and sensing properties of the metal oxide. As far as selectivity is concerned, it is worth to mention that sensor fabrication, and in particular MOX doping and sensor device operating temperature cycling have been suggested as useful means to better tune the device response more specifically toward a given target analyte (Bârsan et al., 1999). One of the limiting factors for the MOX sensing technology is their quite high operating temperature. Baseline drifts can be related indeed to the poor structural stability induced by the high temperatures. The high reactivity of their surfaces, which on one side favors the sensitivity, on the other side can limit their selectivity. Besides, humidity is a common interferent even for these sensing materials and it affects the device response repeatability. To overcome the above difficulties, a huge research effort has been devoted recently to the synthesis of new hybrid materials, to different material processing and device design modifications.

Fig. 7 Effect of crystallite size on the electrical resistance of nanoporous elements of SnO2 at exposure of 800 ppm of H2 or CO in air at an operating temperature of 300  C. Image reprinted with permission provided by Springer Nature and Copyright Clearance Center-License Number 4945940891043.

196

Chemical Sensors: Conductometric Gas Sensors Table 3

ZnO Au-ZnO SnO2 Pt-SnO2 Mo-SnO2 Cu-SnO2 WO3 Pt-WO3 Pd-WO3

Examples of MOX gas sensing properties of unloaded/loaded MOX for different gases. Analyte

Gas concentration

Sensing performance

CO

50 ppm

CH4

1000 ppm

NO

10 ppm

4.2 at 350  C 46.5 at 350  C 0.4 at 500  C 0.7 at 500  C 1.03 at 500  C 1.5 at 500  C 40 at 250  C 70.4 at 150  C 100.3 at 200  C

Data from Wetchakun K, Samerjai T, Tamaekong N, Liewhiran C, Siriwong C, et al. (2011) Semiconducting metal oxides as sensors for environmentally hazardous gases. Sensors and Actuators B: Chemical 160 (1): 580–591.

Fig. 8 Schematic model of spill over process on the surface of doped metal oxide according to a sensitization mechanism that can be electronic or chemical.

Nonoxide materials The second class of conductometric sensing materials is characterized by lower operating temperatures. In principle, this class of innovative materials could reveal to be particularly interesting also in view of stricter energy consumption requirements such as those envisaged in applications related to the Internet of Things scenario. This group of sensing materials mainly comprises polymers, carbon nanotubes and 2D materials.

Polymers In a polymer the adsorption step (a) is described by the partition coefficient (i.e., the ratio of concentrations of the gas in the two phases at equilibrium), which is a function of the analyte partial pressure at constant temperature and the gas diffusion (b) is proportional directly to the available surface area and inversely to the thickness of the polymer layer. The degree of packing of the polymer chains has a noticeable effect on the diffusion rates. In fact, polymer layers with open crystal structures usually allow higher diffusion rates with respect to closely packed crystal structures. In the case of change of the crystal structure during the absorption process, as it is for swelling phenomena, deviations from the Fickian diffusion model can be observed (Hunter et al., 2020). The first polymers investigated in conductometric gas sensing were the insulating polymers, mostly because of the huge variety of chemical structures available. The addition with conducting fillers, such as metals or carbon materials, was the channel to obtain the necessary electrical conductivity. The strength point of this sensing technology relies on the low cost and in the wide range of commercially available polymers that can be used to develop conductive polymer composites (CPCs). The scientific literature shows indeed a great deal of examples of polymers used for the development of gas sensors: polystyrene, polyvinylidenefluoride, polyvinyl acetate, polymethylmethacrylate, polysulfone, polyethyleneoxide, just to name a few. These CPC sensors have demonstrated their effectiveness for vapor sensing especially when set in arrays with sensing elements prepared from different polymers (Chen et al., 2020; Ryan et al., 2004; Craven et al., 1996; De Vito et al., 2008). The CPC films are usually fabricated by simple solution-based processes. The conductive filler, such as a carbon black, is dispersed into a solution containing dissolved insulating polymer to be then dispensed onto a substrate. After the solvent evaporation, the carbon black forms conductive pathways inside the polymer matrix making the films electronically conducting. The

Chemical Sensors: Conductometric Gas Sensors

197

selectivity to different gas analytes results from the specific interactions of the insulating polymer with the gases in the surrounding environment while the electrical response arises from the conductive filler paths modification. The adsorption of gaseous analytes onto the polymer matrix causes in fact the film to swell, disrupting conductive pathways by pushing conductive particles apart, as it is schematically shown in Fig. 8. As a result, the film electronic resistance increases. The ability to swell of the CPC layers depends on the partition coefficient of each analyte. Therefore, polymer selection greatly depends on the suite of target compounds. A properly chosen set of polymers can distinguish compounds of very different chemical structures as well as compounds of very similar structures, such as isomers or even enantiomers when the set includes chiral polymers (Ryan and Lewis, 2001). In the last years, nanomaterials such as carbon nanotubes (CNT) or graphene have been incorporated in polymeric nanocomposites with excellent results (Sun et al., 2013) (Fig. 9). Another class of polymers widely employed in gas sensing is that of intrinsically conductive polymers (CPs). Here the conduction is due to extensive conjugation of alternate double and single bonds in the backbone of the polymer (Heeger, 2001). The delocalization of p-electrons across a few atoms of the polymer backbone provides the means for charge mobility along the polymer chain and this confers to CPs in their pristine state, a semiconducting character. The discovery, in 1977, that the conductivity can be increased up to metallic levels by doping processes, not only was worth being awarded the Nobel Prize for Chemistry in 2000 but also produced a considerable interest in CPs (Shirakawa et al., 1977). The doping of a CP involves the oxidation (removal of some delocalized electrons) or reduction (addition of electrons) of the p-electronic system which makes the CPs p-type and ntype semiconducting materials, respectively. Thick or thin films of CPs deposited between two narrow spaced electrodes is the normal configuration for conductometric gas sensors based on these materials. There are several techniques available for the fabrication of CP films, among which electrospinning, interfacial chemical polymerization and template directed chemical and electrochemical polymerization are the most frequently used. Conductivity of chemiresistor devices based on CPs is assessed by applying a small DC voltage or a constant current (AC or DC) between the electrodes and by measuring the resulting current/voltage. The exposure of the layer to a gas analyte can change its electrical resistance with variations that depend on the nature of sensor materials and on the nature of gas molecules. The interactions mechanisms feasible with this kind of material are multiple indeed. CPs chemiresistors can respond to volatile compounds by conformational changes to the backbone structure, such as swelling or alignment of the chains, or by modifications in the charge carriers. Conformational changes to the polymer backbone structure induce different electrical resistance changes: the swelling of the polymer backbone generally increases the average electron hopping distance between the chains, thus increasing the resistance; conversely, chain alignment increases chain crystallinity, which decreases the resistance. When the gas molecules interact with the p-electrons of the polymer backbone CPs can act as electron donor or acceptor thereby changing the density of charge carriers. If a p-type CP donates electrons to the gas, its conductivity increases. Conversely, when the same CP acts as an electron acceptor, its conductivity decreases. The interacting gas molecules can also change the mobility of the charge carriers thereby modifying the conductivity. These mechanisms enable a wide range of gases and volatile substances to stimulate a sensing response. In addition, regardless of the interaction mechanism, the response of CPs chemiresistors is generally rapid and reversible at room temperature. The first polymers used in the fabrication of conductometric gas sensors were polypyrrole (PPy), polyaniline (PANI), polythiophene (PT), poly (3,4-ethylenedioxythiophene) (PEDOT) and their derivatives (Hunter et al., 2020). These polymers have shown significant sensitivity toward ppm concentrations of reducing compounds (such as NH3, CO, CH4, H2, H2S, acetone, ethanol) as well as toward oxidizing analytes (such as NOx, CO2, SO2, O2, O3), with response times that generally range in the tens of seconds. A summary of the sensing performance of these materials is reported in the following Table 4 (Wong et al.,

Target analyte Air

Air

Substrate

CPC

Electrode

Rg Electrode

CPC

Electrode

Electrode

Ra