Functionalized Inorganic Fluorides: Synthesis, Characterization and Properties of Nanostructured Solids [1 ed.] 0470740507, 9780470740507

Citation preview

Functionalized Inorganic Fluorides

Functionalized Inorganic Fluorides Synthesis, Characterization & Properties of Nanostructured Solids

Edited by ALAIN TRESSAUD Research Director CNRS (Emeritus), ICMCB–CNRS, Bordeaux University, France

A John Wiley and Sons, Ltd., Publication

This edition first published 2010  2010 John Wiley & Sons, Ltd Registered office John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, United Kingdom For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com. The right of the author to be identified as the author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book. This publication is designed to provide accurate and authoritative information in regard to the subject matter covered. It is sold on the understanding that the publisher is not engaged in rendering professional services. If professional advice or other expert assistance is required, the services of a competent professional should be sought. The publisher and the author make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties, including without limitation any implied warranties of fitness for a particular purpose. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for every situation. In view of ongoing research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. The fact that an organization or Website is referred to in this work as a citation and/or a potential source of further information does not mean that the author or the publisher endorses the information the organization or Website may provide or recommendations it may make. Further, readers should be aware that Internet Websites listed in this work may have changed or disappeared between when this work was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising herefrom. Library of Congress Cataloging-in-Publication Data Functionalized inorganic fluorides: synthesis, characterization & properties of nanostructured solids / edited by Alain Tressaud. p. cm. Includes bibliographical references and index. ISBN 978-0-470-74050-7 (pbk.) 1. Fluorides. I. Tressaud, Alain. QD181.F1F77 2010 5460 .731—dc22 2009052139 A catalogue record for this book is available from the British Library. ISBN: 978-0-470-74050-7 (Cloth) Set in 10/12pt Times by Integra Software Services Pvt. Ltd., Pondicherry, India Printed and bound in Great Britain by CPI Antony Rowe, Chippenham, Wiltshire. Cover images from left to right: Projection along [001] of the ITQ-33 zeolite structure showing the 18-MRs windows (Chapter 16); Schematic morphology of oxyfluoride glass-ceramics formed by spinodal decomposition (Chapter 9); Crystal structure of La2CuO3.6F0.8 [The Cu cations are situated in octahedra; the La cations are shown as large spheres; the F anions are shown as small spheres] (Chapter 13)

Contents

Preface List of Contributors 1

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties Erhard Kemnitz, Gudrun Scholz and Stephan Ru¨diger

1

1.1

1

Introduction 1.1.1 Sol-Gel Syntheses of Oxides – An Intensively Studied and Widely Used Process 1.1.2 Sol-Gel Syntheses of Metal Fluorides – Overview of Methods 1.2 Fluorolytic Sol-Gel Synthesis 1.2.1 Mechanism and Properties 1.2.2 Insight into Mechanism by Analytical Methods 1.2.3 Exploring Properties 1.2.4 Possible Fields of Application References 2

xvii xxi

1 2 4 5 8 27 29 35

Microwave-Assisted Route Towards Fluorinated Nanomaterials Damien Dambournet, Alain Demourgues and Alain Tressaud

39

2.1 2.2

39 40 40 40 41 41 42 42

2.3

Introduction Introduction to Microwave Synthesis 2.2.1 A Brief History 2.2.2 Mechanisms to Generate Heat 2.2.3 Advantages of Microwave Synthesis 2.2.4 Examples of Microwave Experiments Preparation of Nanosized Metal Fluorides 2.3.1 Aluminium-based Fluoride Materials 2.3.2 Microwave-assisted Synthesis of Transition Metal Oxy-Hydroxy-Fluorides

61

vi

3

Contents

2.4 Concluding Remarks Acknowledgements References

64 64 65

High Surface Area Metal Fluorides as Catalysts Erhard Kemnitz and Stephan Ru¨diger

69

3.1 3.2 3.3 3.4 3.5 3.6

69 71 74 78 84 88 90 94 95 97

Introduction High Surface Area Aluminium Fluoride as Catalyst Host-Guest Metal Fluoride Systems Hydroxy(oxo)fluorides as Bi-acidic Catalysts Oxidation Catalysis Metal Fluoride Supported Noble Metal Catalysts 3.6.1 Hydrodechlorination of Monochlorodifluoromethane 3.6.2 Hydrodechlorination of Dichloroacetic Acid (DCA) 3.6.3 Suzuki Coupling References

4

Investigation of Surface Acidity using a Range of Probe Molecules Alexandre Vimont, Marco Daturi and John M. Winfield

101

4.1

101 102

Introduction 4.1.1 Setting the Scene: Metal Fluorides versus Metal Oxides 4.1.2 Some Examples of the Application of FTIR Spectroscopy to the Study of Surface Acidity in Metal Oxides 4.1.3 A Preview 4.2 Characterization of Acidity on a Surface: Contrasts with Molecular Fluorides 4.2.1 Molecular Brønsted and Molecular Lewis Acids 4.2.2 A Possible Benchmark for Solid Metal Fluoride, Lewis Acids: Aluminium Chlorofluoride 4.3 Experimental Methodology 4.3.1 FTIR Spectroscopy 4.3.2 Characteristic Reactions and the Detection of Adsorbed Species by a Radiotracer Method 4.4 Experimental Studies of Surface Acidity 4.4.1 Using FTIR Spectroscopy 4.4.2 Using HCl as a Probe with Detection via [36Cl]-Labelling 4.4.3 Metal Fluoride Surfaces that Contain Surface Hydroxyl Groups: Aluminium Hydroxy Fluorides with the Hexagonal Tungsten Bronze Structure 4.4.4 Possible Geometries for HCl Adsorbed at Metal Fluoride Surfaces: Relation to Oxide and Oxyfluoride Surfaces 4.5 Conclusions References

103 107 108 108 109 110 110 112 117 118 123

129 135 136 137

Contents

5

Probing Short and Medium Range Order in Al-based Fluorides using High Resolution Solid State Nuclear Magnetic Resonance and Parameter Modelling Christophe Legein, Monique Body, Jean-Yves Buzare´, Charlotte Martineau and Gilles Silly 5.1 5.2

Introduction High Resolution NMR Techniques 5.2.1 Fast MAS and High Magnetic Field 27 Al NMR 5.2.2 5.2.3 High Resolution Correlation NMR Techniques 5.3 Application to Functionalized Al-Based Fluorides with Catalytic Properties 5.3.1 Crystalline Aluminium Fluoride Phases 19 F Isotropic Chemical Shift Scale in Octahedral Aluminium 5.3.2 Environments with Oxygen and Fluorine in the First Coordination Sphere 5.3.3 Fluorinated Aluminas and Zeolites, HS AlF3 5.3.4 Aluminium Chlorofluoride and Bromofluoride 5.3.5 Pentahedral and Tetrahedral Aluminium Fluoride Species 5.3.6 Nanostructured Aluminium Hydroxyfluorides and Aluminium Fluoride Hydrate with Cationic Vacancies 5.3.7 iso Scale for 27Al and 19F in Octahedral Aluminium Environments with Hydroxyl and Fluorine in the First Coordination Sphere 5.4 Alkali and Alkaline-earth Fluoroaluminates: Model Compounds for Modelling of NMR Parameters 19 F NMR Line Assignments 5.4.1 27 Al Site assignments, Structural and Electronic 5.4.2 Characterizations 5.5 Conclusion References

6

vii

141

141 142 142 145 148 153 153

153 157 158 158

159

160 160 161 164 167 168

Predictive Modelling of Aluminium Fluoride Surfaces Christine L. Bailey, Sanghamitra Mukhopadhyay, Adrian Wander, Barry Searle and Nicholas Harrison

175

6.1 6.2

175 176 176 177 178 179 180 180 180

6.3

Introduction Methodology 6.2.1 Density Functional Theory 6.2.2 Surface Free Energies 6.2.3 Molecular Adsorption 6.2.4 Kinetic Monte Carlo Simulations Geometric Structure of  and -AlF3 6.3.1 Bulk Phases 6.3.2 Surfaces

viii

Contents

6.4 6.5

Characterization of AlF3 Surfaces Surface Composition under Reaction Conditions 6.5.1 The -AlF3–x (01–12) Termination 6.5.2 The -AlF3 (0001) Termination 6.6 Characterization of Hydroxylated Surfaces 6.7 Surface Catalysis 6.7.1 Molecular Adsorption 6.7.2 Reaction Mechanisms and Barriers 6.7.3 Analysing the Kinetics of the Reaction 6.8 Conclusions Acknowledgements References

7

Inorganic Fluoride Materials from Solvay Fluor and their Industrial Applications Placido Garcia Juan, Hans-Walter Swidersky, Thomas Schwarze and Johannes Eicher 7.1 7.2

7.3

7.4 7.5

7.6 7.7 7.8 7.9 7.10 7.11 7.12 7.13 7.14 7.15 7.16 7.17

Introduction Hydrogen Fluoride 7.2.1 Anhydrous Hydrogen Fluoride, AHF 7.2.2 Hydrofluoric Acid Elemental Fluorine, F2 7.3.1 Fluorination of Plastic Fuel Tanks 7.3.2 Finishing of Plastic Surfaces 7.3.3 F2 Mixtures as CVD-chamber Cleaning Gas Iodine Pentafluoride, IF5 Sulfur Hexafluoride, SF6 7.5.1 SF6 as Insulating Gas for Electrical Equipment 7.5.2 SF6 Applications in Metallurgy Ammonium Bifluoride, NH4HF2 Potassium Fluorometalates, KZnF3 and K2SiF6 Cryolite and Related Hexafluoroaluminates, Na3AlF6, Li3AlF6, K3AlF6 Potassium Fluoroborate, KBF4 Fluoboric Acid, HBF4 Barium Fluoride, BaF2 Synthetic Calcium Fluoride, CaF2 Sodium Fluoride, NaF Sodium Bifluoride, NaHF2 Potassium Bifluoride, KHF2 Potassium Fluoroaluminate, KAlF4 Fluoroaluminate Fluxes in Aluminium Brazing 7.17.1 Flux Composition 7.17.2 Flux and HF 7.17.3 Flux Particle Size

185 188 189 192 193 196 197 198 200 201 203 203

205

205 205 206 206 207 207 207 208 208 209 209 209 210 210 211 212 212 213 213 213 213 214 214 214 214 216 217

Contents

8

7.17.4 Flux Melting Range 7.17.5 Current Status of Aluminium Brazing Technology 7.17.6 Cleaning and Flux Application 7.17.7 Wet Flux Application 7.17.8 Dry/Electrostatic Flux Application 7.17.9 Post Braze Flux Residue 7.17.10 Filler Metal Alloys 7.17.11 Flux Precoated Brazing Sheet/Components 7.17.12 Clad-less Brazing 7.17.13 Furnace Conditions 7.18 Summary References

219 220 221 221 222 222 222 223 223 224 224 225

New Nanostructured Fluorocompounds as UV Absorbers Alain Demourgues, Laetitia Sronek and Nicolas Penin

229

8.1 8.2

229 231 231 232

Introduction Synthesis of Tetravalent Ce and Ti-based Oxyfluorides 8.2.1 Preparation of Ce-Ca-based Oxyfluorides 8.2.2 Preparation of Ti-based Oxyfluorides 8.3 Chemical Compositions and Structural Features of Ce and Ti-based Oxyfluorides 8.3.1 Elemental Analysis 8.3.2 Magnetic Measurements 8.3.3 About the Chemical Composition of Ce1xCaxO2x and Ce1xCaxO2xy/2Fy Series 8.3.4 About the Structure and Local Environment of Fluorine in Ce1xCaxO2xy/2Fy Series 8.3.5 Composition and Structure of Ti-based Hydroxyfluoride 8.4 UV Shielding Properties of Divided Oxyfluorides 8.4.1 The Ce-Ca-based Oxyfluorides Series and UV-shielding Properties 8.4.2 Ti Hydroxyfluoride and UV-shielding Properties 8.5 Conclusion Acknowledgement References 9

ix

233 233 233 234 237 252 263 264 266 267 269 269

Oxyfluoride Transparent Glass Ceramics Michel Mortier and Ge´raldine Dantelle

273

9.1 9.2

273 274 275 277 279 281 281

9.3

Introduction Synthesis 9.2.1 Synthesis by Glass Devitrification 9.2.2 Transparency Different Systems 9.3.1 Glass-Ceramics with CaF2 as their Crystalline Phase 9.3.2 Glass-Ceramics with -PbF2 as their Crystalline Phase

x

Contents

9.3.3 Glass-Ceramics with CdF2/PbF2 as their Crystalline Phase 9.3.4 Glass-Ceramics with LaF3 as their Crystalline Phase 9.4 Thermal Characterization 9.4.1 Kinetics of Phase-change/Devitrification 9.4.2 Thakur’s Method 9.5 Morphology of the Separated Phases 9.6 Optical Properties of Glass-Ceramics 9.6.1 Influence of the Devitrification on the Spectroscopic Properties of Ln3þ 9.6.2 Effect of High Local Ln3þ Concentration in Crystallites 9.6.3 Comparison of the Optical Properties of Glass-Ceramics and Single-Crystals 9.6.4 Multi-doped Glass-Ceramics 9.7 Conclusion References 10

11

Sol-Gel Route to Inorganic Fluoride Nanomaterials with Optical Properties Shinobu Fujihara

281 282 282 288 288 289 293 293 295 297 299 301 302

307

10.1 10.2

Introduction Principles of a Sol-Gel Method 10.2.1 Metal Oxide Materials 10.2.2 Metal Fluoride Materials 10.3 Fluorinating Reagents and Method of Fluorination 10.4 Control of Shapes and Microstructures 10.5 Optical Properties 10.5.1 Low Refractive Index and Anti-Reflection Effect 10.5.2 Luminescence 10.6 Concluding Remarks References

307 308 308 308 309 313 317 317 322 326 326

Fluoride Glasses and Planar Optical Waveguides Brigitte Boulard

331

11.1 11.2

331 332 333 334 336 338 340 341 342 344 344

Introduction Rare Earth in Fluoride Glasses 11.2.1 Fundamentals 11.2.2 Applications: Laser and Optical Amplifiers 11.3 Fabrication of Waveguides: A Review 11.4 Performance of Active Waveguides 11.4.1 Optical Amplifier 11.4.2 Lasers 11.5 Fluoride Transparent Glass Ceramics: An Emerging Material 11.6 Conclusion References

Contents

12

13

xi

Polyanion Condensation in Inorganic and Hybrid Fluoroaluminates Karim Adil, Amandine Cadiau, Annie He´mon-Ribaud, Marc Leblanc and Vincent Maisonneuve

347

12.1 12.2 12.3

Introduction Synthesis Extended Finite Polyanions (0D) 12.3.1 Isolated AlF4 Tetrahedra 12.3.2 Isolated AlF6 Octahedra 12.3.3 Al2F11 Dimers 12.3.4 Al3F16 Trimers 12.3.5 Al2F10 Dimers 12.3.6 Al4F20 Tetramers 12.3.7 Al4F18 Tetramers 12.3.8 Al5F26 Pentamers 12.3.9 Al7F30 Heptamers 12.3.10 Al8F35 Octamers 12.3.11 Mixed Polyanions 12.4 1D Networks 12.4.1 AlF5 Chains 12.4.2 Al2F9 Chains 12.4.3 Al7F29 Chains 12.4.4 AlF4 Chains 12.4.5 Mixed Polyanions and/or Chains 12.5 2D Networks 12.5.1 Al3F14 Layers 12.5.2 AlF4 Layers 12.5.3 Al2F7 Layers 12.5.4 Al5F17 Layers 12.5.5 Al3F10 Layers 12.6 3D Networks 12.6.1 Al7F33 Network 12.6.2 Al2F9 Network 12.6.3 AlF3 Network 12.7 Evolution of the Condensation of Inorganic Polyanions 12.7.1 Influence of Amine and Aluminum Concentrations 12.7.2 Temperature Acknowledgements Supplementary Materials References

347 348 350 350 350 353 353 353 354 354 355 355 356 356 358 358 359 360 360 361 365 365 365 366 366 368 368 368 368 369 372 372 374 376 376 376

Synthesis, Structure and Superconducting/Magnetic Properties of Cu- and Mn-based Oxyfluorides Evgeny V. Antipov and Artem M. Abakumov

383

13.1 13.2

383 384

Introduction Chemical Aspects of Fluorination of Complex Oxides

xii

Contents

13.3

Structural Aspects of Fluorination of Complex Cuprates and Superconducting Properties 13.3.1 Electron Doped Superconductors: Heterovalent Replacement 1O2 ! 1F 13.3.2 Hole Doped Superconductors: Fluorine Insertion into Vacant Anion Sites 13.3.3 Structural Rearrangements in Fluorinated Cuprates 13.3.4 Fluorination of Nonsuperconducting Cuprates 13.4 Fluorination of Manganites 13.5 Conclusions References 14

Doping Influence on the Defect Structure and Ionic Conductivity of Fluorine-containing Phases Elena I. Ardashnikova, Vladimir A. Prituzhalov and Ilya B. Kutsenok 14.1 14.2

14.3

14.4 14.5

14.6

14.7

14.8

Introduction Influence of Oxygen Ions on Fluoride Properties 14.2.1 Pyrohydrolysis 14.2.2 Heterovalent Oxygen Substitution for Fluoride Ions 14.2.3 Ionic Conductivity of Oxyfluoride Cation Doping of Fluorides 14.3.1 Isovalent Replacement in the Cation Sublattice 14.3.2 Heterovalent Replacement in the Cation Sublattice Active Lone Electron Pair of Cations and Ionic Conductivity Peculiarities of the Defect Structure of Nonstoichiometric Fluorite-like Phases 14.5.1 Fluorite Structure 14.5.2 Defect Clusters 14.5.3 Ordered Fluorite-like Phases 14.5.4 Phase Diagrams Ionic Transfer in Fluorite-like Phases 14.6.1 Defect Region Model 14.6.2 Nonstoichiometric Fluorites as Examples of Nanostructured Materials Peculiarities of the Defect Structure of Nonstoichiometric Tysonite-like Phases 14.7.1 Tysonite Structure, Tysonite Modifications and Anion Defects 14.7.2 Ordered Tysonite-like Phases Ionic Transfer in Tysonite-like Phases 14.8.1 Fluoride Ions’ Migration Paths in the LaF3 Structure 14.8.2 Temperature Dependences of Ionic Conductivity and Anion Defect Positions 14.8.3 Concentration Dependences of Ionic Conductivity in Tysonite-like Solid Solutions

388 389 390 398 408 411 415 416

423 423 427 427 428 429 431 431 432 432 435 435 435 439 441 441 443 447 449 449 454 454 455 457 459

Contents

15

14.9 Conclusions References

462 462

Hybrid Intercalation Compounds Containing Perfluoroalkyl Groups Yoshiaki Matsuo

469

15.1 15.2

469

Introduction Preparation and Properties of Intercalation Compounds Containing Perfluoroalkyl Groups 15.2.1 Preparation 15.2.2 Exfoliation and Film Preparation 15.2.3 Introduction of Photofunctional Molecules 15.3 Photophysical and Photochemical Properties of Dyes in Intercalation Compounds Containing Perfluoroalkyl Groups 15.3.1 Microenvironment Estimated by using Probe Molecules Showing Photophysical Responses 15.3.2 Photophysical Properties 15.3.3 Photochemical Properties 15.4 Conclusion and Future Perspectives References 16

17

xiii

The Fluoride Route: A Good Opportunity for the Preparation of 2D and 3D Inorganic Microporous Frameworks Jean-Louis Paillaud, Philippe Caullet, Jocelyne Brendle´, Ange´lique Simon-Masseron and Joe¨l Patarin

471 471 475 476 478 478 480 482 484 484

489

16.1 16.2 16.3 16.4 16.5

Introduction Silica-based Microporous Materials Germanium-based Microporous Materials Phosphate-based Microporous Materials Synthetic Clays 16.5.1 Semi-Synthesis 16.5.2 Solid State Synthesis 16.5.3 Hydrothermal Synthesis 16.6 Conclusion References

489 490 499 504 506 507 508 509 510 511

Access to Highly Fluorinated Silica by Direct F2 Fluorination Alain Demourgues, Emilie Lataste, Etienne Durand and Alain Tressaud

519

17.1 17.2 17.3

519 520

17.4

Introduction Mesoporous Silica and Fluorination Procedures About the Chemical Composition and Morphology of Highly Fluorinated Silica FTIR Analysis 17.4.1 About the Content and Nature of OH/Water Groups in Highly Fluorinated Silica

521 523 523

xiv

Contents

17.4.2 17.4.3

FTIR Bands Related to Si-F Bonds Correlation between Silanol Groups on Mesoporous Silica and Grafted Fluorine on Highly Fluorinated Silica 17.5 Thermal Stability and Water Affinity of Highly Fluorinated Silica 17.6 Nuclear Magnetic Resonance (NMR) Investigations 17.6.1 Local Environments in Highly Fluorinated Silica through NMR Experiments 17.6.2 Effect of Fluorination on the Nuclei Environments 17.7 Conclusions on the F2-gas Fluorination Mechanism of Mesoporous Silica Acknowledgements References 18

Preparation and Properties of Rare-earth-Containing Oxide Fluoride Glasses Susumu Yonezawa, Jae-ho Kim and Masayuki Takashima 18.1 18.2

Introduction Preparation and Basic Characteristics of Oxide Fluoride Glasses Containing LnF3 18.2.1 Preparation of Oxide Fluoride Glasses Containing LnF3 18.2.2 Density and Refractive Index 18.2.3 Glass Transition Temperature 18.3 Optical and Magnetic Properties of LnF3-BaF2-AlF3-GeO2 (SiO2) Glasses 18.3.1 Optical Properties of HoF3-BaF2-AlF3-GeO2 Glasses 18.3.2 Optical Properties of CeF3-BaF2-AlF3-SiO2 Glasses 18.3.3 Optical Properties of the Glasses Co-doped with TbF3 and SmF3 18.3.4 Magnetic Property of TbF3 Containing Oxide Fluoride Glasses 18.4 Conclusion References

19

Switchable Hydrophobic-hydrophilic Fluorinated Layer for Offset Processing Alain Tressaud, Christine Labruge`re, Etienne Durand 19.1 19.2 19.3

19.4

Introduction The Principles of the Lithographic Printing Process Experimental Part 19.3.1 Fluorination by Cold rf Plasmas 19.3.2 Wettability Measurements 19.3.3 Surface Analyses Various Types of Surface Modifications using Fluorinated rf Plasmas 19.4.1 Reactive Etching of Porous Alumina using CF4-Plasma Treatment

526 527 530 533 534 534 540 541 542

545 545 546 546 552 553 555 555 557 564 566 568 569

571 571 572 573 573 574 574 575 575

Contents

Switchable Hydrophilic/Hydrophobic Fluorocarbon Layer Obtained on Porous Alumina using c-C4F8 Plasma Treatment 19.5 Comparison of Surface Modifications of Porous Alumina using Various Fluorinated Media: CF4, C3F8 and c-C4F8 19.6 Conclusion Acknowledgements References

xv

19.4.2

Index

578 580 581 582 582 583

Preface

Fluorides and fluorinated materials affect various aspects of modern life. The strategic importance of fluoride materials, and the use of adapted fluorination surface treatments, concern many research fields and applications in areas such as energy production, microelectronics and photonics, catalysis, colour pigments, textiles, cosmetics, plastics, domestic wares, automotive technology and building. Among the issues with which they are concerned [1–4] are: • the historical importance of fluoride fluxes in the production of metals, in particular aluminium; • the critical place of fluorine and fluorides in conversion energy processes – for example components of Li-ion batteries and fuel cells, enrichment of 235U through uranium hexafluoride for nuclear energy; • the etching of silicon wafers for microelectronics; • the technical revolution of fluoropolymers and fluoride coatings, for example Teflon and fluorinated plastics, waterproof clothes, biomaterials for cardiovascular or retinal surgeries, kitchen wares, and so forth; • the beneficial influence of fluoride on dental caries; • the dominant use of fluorinated molecules in agrochemistry and phytosanitary products; • the dramatic increase of fluorine-containing molecules for medicine and pharmacy, as efficient imaging products, as dental composites for cariostatic improvement, and so forth; • the use of 18F-labelled molecules in positron emission tomography (PET) for early diagnosis of cancer and Alzheimer’s disease. In the case of inorganic fluorinated solids, numerous improvements have recently been achieved through the elaboration and functionalization of the materials on a nanometric scale. The present book covers several classes of nanostructured and functionalized inorganic fluorides, oxide-fluorides, hybrids, mesoporous materials and fluorinated oxides such as silica and alumina. The morphologies concerned range from powders or glassceramics to thin layers and coatings whereas the applications involved include catalysts, inorganic charges, superconductors, ionic conductors, ultaviolet (UV) absorbers, phosphors, materials for integrated optics, and so forth. Several books have been devoted to the reactivity of carbon-based materials with fluorine (carbon fibres, fullerene, carbon nanotubes, etc) [1,2,5,6], so these types of materials will not be treated in the present book.

xviii

Preface

The book arose from discussions that took place during the FUNFLUOS project (2004–2008), carried out within the Sixth European Framework Programme. This project involved about ten groups from Germany, France, Slovenia and the UK, all aimed at the synthesis and characterization of fluorinated materials with properties tailored for specific applications. The topics appearing in the book range from new synthesis routes to physical-chemical characterizations. They address important properties of these materials, including morphology, structure, thermal stability, superconductivity, magnetism, spectroscopic and optical behaviour. Detailed ab initio investigations and simulations provide a comparison with experimental results, and potential applications of the final products are also proposed. In the first section, two innovative routes toward nanoscaled metal fluorides and hydroxyfluorides are presented: the fluorolytic sol-gel synthesis by E. Kemnitz et al. and the microwave-assisted route by D. Dambournet et al. In a second section, several physical-chemical characterizations are developed in order to understand the mechanisms that are responsible for the improvement of the properties of these materials: investigation of the main characteristics of high-surface-area aluminium fluorides as catalysts by E. Kemnitz and S. Ru¨diger; determination of surface acidities (Lewis and Brønsted types) using a large range of probe molecules, by A. Vimont et al.; a better knowledge of the environment of the different nuclei using high-resolution solid-state nuclear magnetic resonance (NMR) by C. Legein et al. The theoretical investigation of these topics is highlighted by the predictive modelling of aluminium fluoride surfaces by C. Bailey et al., which allows a better understanding of the underlying processes at the molecular and nano levels. An example of industrial application of the inorganic fluorides is given by P. Garcia Juan et al. In the following section, some examples of outstanding optical properties of nanostructured fluorides are proposed: nanostructured fluorocompounds as UV absorbers, by A. Demourgues et al.; transparent oxyfluoride glass-ceramics by M. Mortier and G. Dantelle; luminescent and antireflective coating of (oxy)fluorinated materials obtained by the sol-gel technique, by S. Fujihara; planar optical waveguides based on fluoride glasses, by B. Boulard. Hybrids, composites and mesoporous fluorides are original materials with great potential and the interesting nature of such materials is illustrated in the next section by the chapters on polyanion condensation in inorganic-organic hybrid fluorides, by K. Adil et al.; superconducting/magnetic properties of Cuand Mn-based oxyfluorides, by E. Antipov and A. Abakumov; ionic conductivity of fluoride-containing phases by E. Ardashnikova et al.; intercalation in hybrid compounds containing perfluoroalkyl groups, by Y. Matsuo. The two following chapters deal with the synthesis of microporous frameworks using the fluoride and F2-gas routes, respectively. The examples concern either compounds based on silica, germanium, phosphates and clays, by J. L. Paillaud et al., or highly fluorinated silica, by A. Demourgues et al. The optical and magnetic properties of oxyfluoride glasses based on rare-earth elements are illustrated by S. Yonezawa et al. Finally the chapter by A. Tressaud et al. describes the use of surface fluorination of porous alumina for applications in offset technology. A very wide range of materials, properties, and applications have therefore been gathered in this book, which covers various new fields in which inorganic fluorides are part of the innovating process. Among the information that can bring answers to some crucial questions in materials science, we can quote new synthesis routes towards more

Preface

xix

efficient and less aggressive catalysts, protection against harmful UV radiation, new integrated lasers and optical amplifiers, antireflective coatings, solid-state ionic conductors, highly hydrophobic silica and switchable coatings for offset technology. Erhard Kemnitz and Alain Tressaud Berlin and Bordeaux September 2009

References [1] Advanced Inorganic Fluorides, T. Nakajima, B. Zemva, A. Tressaud (Eds), Elsevier, Amsterdam (2000). [2] Fluorinated Materials for Energy Storage, T. Nakajima, H. Groult (Eds), Elsevier, Amsterdam (2005). [3] Fluorine and the Environment, Vol. 1 and Vol. 2, A. Tressaud (Ed.), Elsevier, Amsterdam (2006). [4] Fluorine and Health, A. Tressaud and G. Haufe (Eds), Elsevier, Amsterdam (2008). [5] Graphite Fluorides and Carbon-Fluorine Compounds, T. Nakajima (Ed.), CRC Press, Boca Raton, FL (1991). [6] ‘Fluorofullerenes’, in Dekker Encyclopedia of Nanoscience and Nanotechnology, O. V. Boltalina, S. H. Strauss, 2nd edition, Dekker, New York (2009).

The Funfluos European Network (2004): First row (from left to right): D. Menz, B. Zˇemva, E. Kemnitz (Coordinator), A. Demourgues, A. Tressaud, and J. Winfield. Second row (from left to right): U. Gross, M. Feist (partly hidden), S. Ru¨diger, P. Millet (European Commission), N. Harrison, A. Wander, T. Skapin and S. Schro¨der

List of Contributors Artem M. Abakumov, Department of Chemistry, Moscow State, University, Moscow, Russia Karim Adil, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France Evgeny V. Antipov, Department of Chemistry, Moscow State, University, Moscow, Russia Elena I. Ardashnikova, Department of Chemistry, Moscow State, University, Moscow, Russia Christine L. Bailey, Computational Science and Engineering Department, STFC Daresbury Laboratory, Warrington, Cheshire, UK Monique Body, Laboratoire de Physique de l’Etat Condense´, UMR-CNRS, Universite´ der Maine, Le Mans, France Brigitte Boulard, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France Jocelyne Brendle´, Laboratoire de Mate´riaux a` Porosite´ Controˆle´e, UMR-CNRS, Universite´ de Haute Alsace, Mulhouse, France Jean-Yves Buzare´, Laboratoire de Physique de l’Etat Condense´, UMR-CNRS, Universite´ der Maine, Le Mans, France Amandine Cadiau, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France Philippe Caullet, Laboratoire de Mate´riaux a` Porosite´ Controˆle´e, UMR-CNRS, Universite´ de Haute Alsace, Mulhouse, France Damien Dambournet, Institute of Condensed Matter Chemistry of Bordeaux (ICMCBCNRS), University Bordeaux 1, Pessac, France Ge´raldine Dantelle, Laboratoire de Photonique Quantique et Mole´culaire (LPQM), UMR CNRS, Cachan, France

xxii

List of Contributors

Marco Daturi, ENSICAEN, Universite´ de Caen, CNRS, Caen, France Alain Demourgues, Institute of Condensed Matter Chemistry of Bordeaux (ICMCBCNRS), University Bordeaux 1, Pessac, France Etienne Durand, Institute of Condensed Matter Chemistry of Bordeaux (ICMCB-CNRS), University Bordeaux 1, Pessac, France Johannes Eicher, Solvay Fluor GmbH, Hannover, Germany Shinobu Fujihara, Department of Applied Chemistry, Faculty of Science and Technology, Keio University, Yokohama, Japan Placido Garcia Juan, Solvay Fluor GmbH, Hannover, Germany Nicholas Harrison, Computational Science and Engineering Department, STFC Daresbury, Laboratory, Warrington, Cheshire, UK Department of Chemistry, Imperial College London, London, UK Annie He´mon-Ribaud, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France Erhard Kemnitz, Institute for Chemistry, Humboldt University of Berlin, Berlin, Germany Jae-ho Kim, Graduate School of Engineering, University of Fukui, Fukui, Japan Ilya B. Kutsenok, Department of Chemistry, Moscow State, University, Moscow, Russia Christine Labruge`re, Institute of Condensed Matter Chemistry of Bordeaux (ICMCBCNRS), University Bordeaux 1, Pessac, France Emilie Lataste, Institute of Condensed Matter Chemistry of Bordeaux (ICMCB-CNRS), University Bordeaux 1, Pessac, France Marc Leblanc, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France Christophe Legein, Laboratoire des Oxydes et Fluorures, CNRS, Le Mans, France Vincent Maisonneuve, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France Charlotte Martineau, Laboratoire des Oxydes et Fluorures, UMR CNRS, Le Mans, France, Tectospin, Universite´ de Versailles Saint Quentin en Yvelines, Versailles, France Yoshiaki Matsuo, Department of Materials Science and Chemistry, University of Hyogo, Hyogo, Japan

List of Contributors

xxiii

Michel Mortier, Laboratoire de Chimie de la Matie`re Condense´e de Paris, UMR CNRS, Paris, France Sanghamitra Mukhopadhyay, Department of Chemistry, Imperial College London, London, UK Jean-Louis Paillaud, Laboratoire de Mate´riaux a` Porosite´ Controˆle´e, UMR-CNRS, Universite´ de Haute Alsace, Mulhouse, France Joe¨l Patarin, Laboratoire de Mate´riaux a` Porosite´ Controˆle´e, UMR-CNRS, Universite´ de Haute Alsace, Mulhouse, France Nicolas Penin, Institute of Condensed Matter Chemistry of Bordeaux (ICMCB-CNRS), University Bordeaux 1, Pessac, France Vladimir A. Prituzhalov, Department of Chemistry, Moscow State, University, Moscow, Russia Stephan Ru¨diger, Institute for Chemistry, Humboldt University of Berlin, Berlin, Germany Gudrun Scholz, Institute for Chemistry, Humboldt University of Berlin, Berlin, Germany Thomas Schwarze, Solvay Fluor GmbH, Hannover, Germany Barry Searle, Computational Science and Engineering Department, STFC Daresbury Laboratory, Warrington, Cheshire, UK Gilles Silly, Institut Charles Gerhardt Montpellier, Physicochimie des Mate´riaux De´sordonne´s et Poreux, Universite´ de Montpellier II, Montpellier, France Ange´lique Simon-Masseron, Laboratoire de Mate´riaux a` Porosite´ Controˆle´e, UMRCNRS, Universite´ de Haute Alsace, Mulhouse, France Laetitia Sronek, Institute of Condensed Matter Chemistry of Bordeaux (ICMCB-CNRS), University Bordeaux 1, Pessac, France Hans-Walter Swidersky, Solvay Fluor GmbH, Hannover, Germany Masayuki Takashima, Graduate School of Engineering, University of Fukui, Fukui, Japan Alain Tressaud, Institute of Condensed Matter Chemistry of Bordeaux (ICMCB-CNRS), University Bordeaux 1, Pessac, France

xxiv

List of Contributors

Alexandre Vimont, ENSICAEN, Universite´ de Caen, CNRS, Caen, France Adrian Wander, Computational Science and Engineering Department, STFC Daresbury Laboratory, Warrington, Cheshire, UK John M. Winfield, Department of Chemistry, University of Glasgow, Glasgow, UK Susumu Yonezawa, Graduate School of Engineering, University of Fukui, Fukui, Japan

1 Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties Erhard Kemnitz, Gudrun Scholz and Stephan Ru¨diger Humboldt-Universita¨t zu Berlin, Institut fu¨r Chemie, Brook – Taylor – Str. 2, D – 12489 Berlin, Germany

1.1

Introduction

Sols are dispersions of nanoscopic solid particles in, for example, liquids – i.e., colloidal solutions. The particles can agglomerate forming a three-dimensional network in the presence of large amounts of the liquid thus forming a gel. Inorganic sols are prepared via the sol-gel process, the investigation of which started in the nineteenth century. This process received great impetus from the investigations of Sto¨ber et al. [1], who studied the use of pH adjustment on the size of silica particles prepared via sol-gel hydrolysis of tetraalkoxysilanes. 1.1.1

Sol-Gel Syntheses of Oxides – An Intensively Studied and Widely Used Process

Hydrolysis of alkoxysilanes and later on of metal alkoxides in organic solutions has become an intensively studied and widely used process [2]. The most common products are almost homodispersed nanosized silica or metal oxide particles for, e.g., ceramics or

Functionalized Inorganic Fluorides: Synthesis, Characterization & Properties of Nanostructured Solids Edited by Alain Tressaud  2010 John Wiley & Sons, Ltd

2

Functionalized Inorganic Fluorides

glasses, or the aqueous sols are used to prepare different coatings for, e.g., optical purposes. Optical applications depend on differences in the respective indices of refraction of the coated material and the applied layer. The latter has to be of very uniform and thoroughly adjusted thickness. The sol-gel hydrolysis of alkoxysilanes, the most intensively explored one, basically proceeds in two steps. The first step is the hydrolytic replacement of alkoxy groups, OR, by hydroxyl groups, OH, shown schematically in Equation (1.1) for the first alkoxy group: SiðORÞ4 þ H2 O ! ðRO Þ 3 SiOH þ ROH

(1:1)

Because of their relatively high hydrolytic stability, hydrolysis of alkoxysilanes (1.1) has to be catalysed by Brønstedt acids or bases. In the second step, the primary hydrolysis products undergo condensation reactions under elimination of water (Equation (1.2)) or alcohol (Equation (1.3)). X3 Si-OH þ HO-SiX3 ! X3 Si-O-SiX3 þ H2 OðX ¼ OR; OHÞ

(1:2)

X3 Si-OH þ RO-SiX3 ! X3 Si-O-SiX3 þ ROHðX ¼ OR; OHÞ

(1:3)

As a result tiny particles with a very open structure are formed. The overall process can be controlled by adjusting the reaction conditions. The colloidal solution of these particles, the sol, can be used as such for, e.g., coating or it can be worked up to yield, eventually, nanoscopic oxide particles. However, metal oxide sols obtained in this way always contain a sometimes remarkable organic part. Its separation demands calcination temperatures of at least 623 K in order to convert the ‘precursors’ into pure metal oxide materials. Substituting a certain part of the alkoxidic groups by nonhydrolysable ones, such as alkyl groups in the case of alkoxysilanes or phosphonic acid in the case of metal alkoxides, organically modified oxides, i.e. inorganic-organic hybrid materials, have been prepared.

1.1.2

Sol-Gel Syntheses of Metal Fluorides – Overview of Methods

Selected metal fluorides can, in application-relevant fields, outperform metal oxides and silica. Thus, for instance, magnesium fluoride and aluminium fluoride and, in particular, alkali hexafluoroaluminates have both a lower index of refraction and a much broader spectral range of transparency even than silica, making them very interesting for optical layers. Consequently, several approaches for the preparation of nanoscopic metal fluorides and metal fluoride thin layers have been developed and proposed. Besides physical methods such as milling, laser dispersion or molecular-beam epitaxy, different chemical methods exist. Basically, three approaches can be distinguished: (i) Postfluorination of a metal oxide preformed via sol-gel route [3]. This route is shown schematically in Scheme 1.1. The disadvantages of this route are, to name two, incomplete fluorination of the bulk metal oxides and decrease of surface area in course of the fluorination.

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties M(OR)x + xH2O hydrolysis

MOx/2(gel) bulk

3

MOx/2 (sol) + xROH coating

MOx/2(gel) film fluorination

fluorination + calcination MFx or MOx/2/MFx glass

MFx or MOx/2 /MFx film

Scheme 1.1 Metal fluoride preparation via post fluorination of sol-gel prepared metal oxides (Reproduced from [4] by permission of Elsevier Publishers)

(ii) Preparation of a precursor containing a metal compound with organically bound fluorine such as trifluoroacetate, which is calcined to decompose the fluoroorganic component under formation of metal fluoride [5]. This route, shown in Scheme 1.2, also starts from metal alkoxides, which are reacted in solution with, e.g., trifluoroacetic acid to form metal trifluoroacetate sol. This can be M(OR)x and/or MOAc + CF3COOH + H2O + org. solvent

metal trifluoroacetate (TFA) precursor solution drying

metal-TFA(gel) bulk

coating

metal-TFA(gel) film heating

heating metal fluoride powder

metal fluoride film

heating at higher temperature

MOx/2 /MFx powder

MOx/2 /MFx film

Scheme 1.2 Metal fluoride preparation via metal fluoroacetate sol-gel formation and following thermal decomposition. (Reproduced from [4] by permission of Elsevier Publishers)

4

Functionalized Inorganic Fluorides

used for coating experiments. The decisive final step is the thermal decomposition of the fluoro-organic constituent, because of which thermolabile materials cannot be coated. Another disadvantage is the probability that oxidic components can be formed as admixtures or oxofluorides. (iii) Fluorolytic sol-gel process as counterpart to the hydrolytic one. The fluorolytic sol-gel route follows rather strictly the ‘classical’ hydrolytic one by reacting metal alkoxides in anhydrous solution with hydrogen fluoride instead of the hydrogen oxide of the ‘classical’ process. Consequently, it results eventually in metal fluorides instead of metal oxides. The fluorolytic sol-gel process, its execution, mechanism, scope as well as properties and possible fields of application of its products are the subjects of this chapter.

1.2

Fluorolytic Sol-Gel Synthesis

Metal alkoxides can be regarded as metal salts of alcohols, where the latter are very weak Brønstedt acids. Acids that are stronger than the respective alcohol can therefore replace alkoxy groups attached to the metal ion under liberation of the alcohol and formation of the metal fluoride according to Equation (1.4). MðORÞn þ x HF ! MðORÞn  x Fx þ x ROH ðM ¼ metal ionÞ

(1:4)

In fact, starting with aluminium isopropoxide [6], a broad range of metal alkoxides have been subjected to a sol-gel-like liquid-phase fluorination with hydrogen fluoride in organic solution [4, 7]. Although Equation (1.4) closely resembled Equation (1.1) there is an important difference in that condensation reactions like those of Equations (1.2) and (1.3) are not possible in the fluorolysis system. On the other hand, the fluorolysis reactions typically result in the formation of a sol-gel. The formation of a gel was already mentioned in the first paper on metal alkoxide fluorolysis, reporting the reaction of aluminium isopropoxide in alcoholic solution with an ethereal solution of hydrogen fluoride [6]. The gel formation is obviously due to an important consequence of the replacement of alkoxy groups by fluoride, i.e., the Lewis acidity of the metal ion increases leading to a strengthening of the interaction between (liberated) alcohol molecules and metal ions. As a result alcohol molecules that can occupy ligand positions might establish a loose net between (partly) fluorinated metal ions resulting eventually in metal fluoride sol or even gels. Surprisingly, attempts to isolate pure AlF3 by drying and calcining the gel were not successful; the product obtained had an understoichiometric amount of fluorine even when the primary reaction has been carried out with an overstoichiometric amount of HF [8]. An additional fluorination of the dried gel under gentle conditions (see below) has proved to be a suitable way to remove the attached organic components resulting in X-ray amorphous, highly Lewis acidic aluminium fluoride with unusual large specific surface area, named HS-AlF3 [9].

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

1.2.1

5

Mechanism and Properties

To gain insight into the fluorolytic sol-gel process, elucidating the mechanism and optimizing both the experimental procedure and the properties of the products, the influence of all adjustable synthesis parameters was tested as well as the process. In particular, the products were analysed by a broad range of analytical, experimental and theoretical methods. Most of these investigations focussed on the synthesis and properties of HS-AlF3, followed by HS-MgF2. The results and consequences of these investigations are presented and discussed below as well as in subsequent chapters. 1.2.1.1

Approach to Mechanism and Resulting Properties

The fluorolytic sol-gel synthesis of metal fluorides is summarized in Scheme 1.3. [M(OR)n]x

+S

[M(OR)n]y /S

+

HF/S’

1 2 M-F-SS’ sol 3 M-F-SS’ wet gel 4 M-F dry gel 5 HS-MFn

Scheme 1.3 Fluorolytic sol-gel synthesis of high surface area metal fluorides (for explanations of the numbers 1 to 5 see text)

The synthesis was investigated in detail, aiming its optimization for HS-AlF3 [8] and also for HS-MgF2 [10, 11]. In short, the metal alkoxide, the structure of which can be very complex [12], is dissolved in an alcohol or another suitable organic solvent (step 1 in Scheme 1.3), so that in the case of aluminium isopropoxide the tetrameric structure is predominantly preserved. A solution of hydrogen fluoride in, e.g., ether or alcohol is added in approximately stoichiometric amounts to the alkoxide solution. Ratios of HF:Al from 2 to 4 are well tolerated, resulting in a clear, translucent sol (step 2 in Scheme 1.3), which more-or-less rapidly becomes a gel, depending on concentration and type of solvent (step 3 in Scheme 1.3). Varying the type of alkoxide, from methoxide to ethoxide, isopropoxide to butoxide, did not markedly affect the outcome of the fluorolytic reaction.The only meaningful criteria for estimation of these and other synthesis parameters were thus the surface area and especially Lewis acidity of the final HS-AlF3.

6

Functionalized Inorganic Fluorides

There is yet another route to the synthesis of metal fluoride sol-gel, exemplified for aluminium and magnesium, namely the direct reaction of the metal with an alcoholic HF solution [13]. Upon drying the sol-gel under mild conditions, at about 343 K under vacuum or freeze drying, or under microwave irradiation, a solid, X-ray amorphous dry gel is formed (step 4 in Scheme 1.3), which contains, in case of the Al-F-system, large amounts of organic material, indicated by a carbon content of about 20 %–30 %. An empirical formula based on elemental analysis for a dry gel prepared from Al(OiPr)3 in iPrOH is AlF2.7[OCH(CH3)2]0.30.7-0.8(CH3)2CHOH. With metal ions of lower Lewis acidity, decisive lower carbon contents were found, e.g. 3 %–7 % C in the Mg-F-system. Obviously, part of the alcohol, which is the predominant constituent of the wet gel, is very tightly attached to the highly Lewis acidic Al3þ ion, as can also be seen from its thermo-analysis (Figure 1.1). The weakly endothermic mass loss of about 24 % up to 473 K can clearly be attributed to the release of solvating iPrOH and the more pronounced smaller one around 495 K to the split off of alkoxide groups. Evaluating thermal analysis data of many different experiments, it became obvious that the mass loss proceeds stepwise. The alcohol content of the steps corresponds to the respective compositions of about AlF3:1ROH after heating at 343 K under vacuum, about AlF3:0.45ROH after heating at 573 K in N2, and about AlF3:0.1ROH after continued heating up to about 600 K [13]. The exothermal peak in Figure 1.1 at 836 K is due to crystallization, i.e. of -AlF3 formation. In order to obtain a still X-ray amorphous aluminium fluoride, the dry gel with its understoichiometric fluorine content has to be freed from its organic constituents under fluorinating conditions. This can be accomplished in a gas-solid reaction at elevated temperatures up to 573 K with vaporized fluorocarbon compounds such as CHClF2 diluted with an inert gas. An aluminium fluoride is obtained, named HS-AlF3, which is still amorphous, has a specific surface area of about 200 m2/g and shows extremely high Lewis acidity (see below). An unwanted consequence of the extreme Lewis acidity is the readily occurring coke formation preventing the use of fluorochloroethane compounds as fluorinating agents. The postfluorination step is essential for HS-AlF3 formation, therefore all parameters have been comprehensively tested, such as type and concentration of the fluorinating agent, flow rate, temperature, and also ageing of the Al-F-sol. Even under optimum conditions with CHClF2 or CH2F2 the formation of black spots or sometimes of ‘channels’ could be observed indicating that at these spots the formation of HS-AlF3 had started, which then subsequently catalysed coke formation. Such side reactions are, for obvious reasons, not possible using HF as fluorinating agent. In addition, the stronger fluorinating HF can be used at lower temperatures and should preserve the amorphous state with its high surface area even better. The latter criterion could only be fulfilled using rather diluted HF, obviously to reduce the otherwise high reaction enthalpy. However, ‘HS-AlF3’ prepared with gaseous HF did not show the expected Lewis acidity. It turned out that HF had behaved as base, which became attached to the strongest Lewis acid sites of the solid, thereby blocking them. Only by additional longer flushing with a stream of inert gas or, even better, of CHClF2 vapour at elevated temperatures (up to 573 K) did the material become the expected strong solid Lewis acid, i.e. HS-AlF3. HS-AlF3 could also be prepared with elemental fluorine using a dry Al-F-gel of low carbon content obtained by microwave heating of an Al-F-sol under autologous pressure followed by microwave-assisted drying.

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties DTA /µV DTG /(%/min)

TG (%) TG 100 90 80

–23.85%

50

10

DTG

a –9.48%

70 DTA 60

–5.65%

472

↑exo

836

415

821

30 373

TG (%) TG 100

473

573 673 Temperature/K

773

–15

m41 m45

b

70

30

m87 (x100) 373

473

573 673 Temperature/K

–5

DTG /(%/min) Ion Current –11∗10/A [2]

80 DTG

40

0

0 –0.50 –1.00 –1.50 –2.00 –2.50 –3.00 –3.50

873

90

50

5

–10

495

40

60

7

773

873

0 –0.50 –1.00 –1.50 –2.00 –2.50 –3.00 –3.50

6.0 5.0 4.0 3.0 2.0 1.0 0

Figure 1.1 TA-MS of AlF2.7[OCH(CH3)2]0.30.70.8(CH3)2CHOH; (a) Thermoanalytical curves in N2 (19.50 mg); (b) Ion current curves for m/z 41 (C3H5þ), m/z 45 (C2H5Oþ), and m/z 87 (C5H11Oþ), indicating the release of propene, i-propanol, and diisopropyl ether, respectively. (Reprinted with permission from [9] Copyright (2005) RSC.)

The results of experimental investigation are in accordance with the following tentative mechanism of the sol-gel fluorination. Upon addition of HF to the alkoxide solution a stepwise replacement of –OR by –F starts, whereby the coordinating alcohol as linking group prevents the formation of a three dimensional purely F-linked crystal. The sol-gel state, almost immediately formed, kinetically prevents the stoichiometric fluorination of the aluminium species. Upon removing the alcohol under gentle conditions the gel structure only partly collapses; alcohol molecules of the first ligand sphere remain obviously attached to Al. When these molecules are removed under appropriate mild conditions there is no crystallization taking place but the disordered X-ray amorphous state connected with an unusual high surface area remains and a part of the Al atoms becomes co-ordinately unsaturated consequently exhibiting very high Lewis acidity (see below). For HS-AlF3 a surface area up to 400 m2/g has been determined by N2 adsorption/desorption experiments. Typical isotherms are shown in Figure 1.2.

Functionalized Inorganic Fluorides Pore volume dV/d logD [cm3/g]

8

Volume adsorbed [cm3/g]

250

200

150

2

1

0 0

100 200 300 Pore diameter [Å]

100

50 adsorption desorption 0 0.0

0.1

0.2

0.3 0.4 0.5 0.6 0.7 Relative Pressure (p/p 0)

0.8

0.9

1.0

Figure 1.2 N2 adsorption/desorption isotherms and pore size distribution of HS-AlF3 (Reprinted with permission from [6] Copyright (2003) Wiley-VCH.)

The tentative mechanism is given in more detail and supported by the analytical investigations discussed in the following section.

1.2.2

Insight into Mechanism by Analytical Methods

Different spectroscopic, microscopic and diffraction methods like IR and Raman spectroscopy, TEM or XRD were applied to characterize educts, intermediates or products of the sol-gel process. For a detailed insight into the mechanism of the fluorolytic sol-gel process, however, the application of NMR spectroscopy is the method of choice. The NMR experiments, both in liquid and in solid state, allow direct observations of local structures and their changes even if the matrices suffer from a loss of lattice periodicity. For the fluorolytic so-gel process both liquid state NMR experiments were realized for the alkoxide solutions, sols and thin gels as well as solid state MAS NMR experiments for the alkoxide, dried alkoxide fluoride gels and high-surface fluorides including 1H, 13C, 27 Al and 19F as sensitive spin probes. The same spin probes along with the use of 1D and 2D NMR experiments allow local structures in liquids and solids to be addressed and directly compared, to follow their changes with a progressive degree of fluorination and finally to derive a possible reaction pathway for this process. Due to the good experimental accessibility of the mentioned spin probes, a detailed study was conducted for the reaction steps ending up with HS-AlF3. Results of these studies are presented briefly in the following section.

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

1.2.2.1

9

Sol-Gel Reaction

1.2.2.1.1 Aluminiumisopropoxide (Al(OiPr)3) as Solid Precursor Compound The basis for an understanding of changes during the fluorination process is the knowledge and assignment of molecular structures existing in solid Al(OiPr)3 and its solution in iPrOH. 27 Al MAS NMR measurements of solid Al(OiPr)3 give unambiguous indications for the existence of two distinguishable aluminium sites in the matrix in agreement with XRD findings [14, 15] (Figure 1.3). A simulation of the spectrum (Figure 1.3) results in typical chemical shift values for AlO4 (di ¼ 61.5 ppm) and AlO6 (di ¼ 1.7 ppm) units [16, 17]. Both the kind and the intensity ratio of AlO4 : AlO6 as 3 : 1 confirm the tetrameric molecular structure (see also Figure 1.5, 1) existing in the tetragonal crystal structure [14, 15].

50

400

200

0

0

(ppm)

–200

–50

–100

–400

(kHz)

Figure 1.3 27Al MAS NMR spectrum of Al(OiPr)3 ( r ¼ 20 kHz; B0 ¼ 9.4 T) (solid line: experiment, dotted line: simulation; insert: central transitions with quadrupolar splitting) (Reprinted with permission from [16] Copyright (2006) Humboldt University.)

In addition to 27Al, the respective 1H-13C CP MAS NMR spectrum allows a distinction to be made not only between CHO and CH3 groups but also between terminal (Al-O) and bridging (Al-O-Al) isopropoxide units. Very narrow line widths in part allow the assignment of 18 distinguishable carbon sites in the NMR spectrum in agreement with the crystal structure (see Figure 1.10 (top)) [14, 15]. According to suggestions made by Abraham [17], terminal (63 ppm) and bridged (66 ppm) CHO – groups are located in the low-field part of the spectrum; terminal CH3 groups are detected in the range between 28–30 ppm and bridging CH3 groups dominate the high-field part of the spectrum (see Figure 1.10 (top)), [16]). 1.2.2.1.2 Aluminiumisopropoxide (Al(OiPr)3) Dissolved in Isopropanol First 1H NMR studies on possible structures of Al(OiPr)3 in different solutions range back to 1963 [18] followed by first 27Al NMR measurements in 1973 [19]. Since that

Functionalized Inorganic Fluorides

10

time aluminium isopropoxide solutions and possible species therein, using CCl4, benzene or toluene as solvents, are subjects of current interest applying different NMR techniques [20–22]. Isopropanol is the standard solvent used for the fluorolytic sol-gel synthesis [4, 6–8], so results of a careful reinvestigation of possible isopropanolic solution are presented in summary. 27 Al NMR spectra of Al(OiPr)3 dissolved in isopropanol are shown in Figure 1.4a,b; the spectrum recorded for the same sample dissolved in diethylether is depicted in Figure 1.4c.

a

b

250

200

150

100 50 (ppm)

0

–50 –100

c

250

200

150

100 50 (ppm)

0

–50 –100

Figure 1.4 27Al NMR spectra of different Al(OiPr)3 – solutions: a and b Al(OiPr)3 in iPrOH recorded with a 400 MHz – spectrometer (a) and a 600 MHz – spectrometer (b), c Al(OiPr)3 in diethylether (400 MHz – spectrometer). For all: solid: experimental spectrum, dashed: simulation and dotted: decomposition. (Reprinted with permission from [23] Copyright (2007) American Chemical Society.)

The spectrum obtained for the etheric solution (Figure 1.4c) indicates the existence of only tetrameric aluminium isopropoxide species in solution (AlO6 (2.5 ppm, 26.7 %), AlO4 (61.8 ppm, 73.3 %), 1 in Figure 1.5) but the situation in isopropanolic solution is much more complex. Applying two different fields (Figure 1.4a,b), both spectra reveal, beside AlO6 and AlO4, the existence of an additional fivefold coordinated aluminium species AlO5 at 32 ppm. They support therewith the existence of trimeric Al(OiPr)3 species (3 in Figure 1.5). Calculated intensities of the spectra support the assumption of further species. Cyclic trimeric species (2 in Figure 1.5) with possible distorted AlO4 polyhedra explain the strong low-field shifted resonance observed at 85 ppm [16, 23].

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties O O Al O

Al O

O

O O Al O O O Al O O 1

11

O

O O O Al O Al O O O Al O O

Al O O O O Al Al O O O

2

3

O

O = OiPr

Figure 1.5 Possible structural units of Al(OiPr)3 in solution. (Reprinted with permission from [23] Copyright (2007) American Chemical Society.)

Moreover, an equilibrium between different trimeric species seems to be feasible, which involves a trimeric species with central sixfold oxygen-coordinated aluminium. The latter may result from the interaction of trimeric species 3 (Figure 1.5) and a solvent molecule as shown in Figure 1.6 [16].

O O Al O

O O

O Al O 4

Al O

O

HOiPr

HOiPr

O Al O

O O

Al O 4'

O O

O Al O O : OiPr

Figure 1.6 Possible equilibria between different trimeric Al(OiPr)3 species in isopropanolic solution. (Reprinted with permission from [16] Copyright (2006) Humboldt University.)

In the appropriate 1H and 13C NMR spectra, no indications are observable that distinguish between tetramers 1 and trimers 2 and 3 (Figure 1.5), respectively. The only possible discrimination between bridging and terminal isopropoxide groups holds for all of the mentioned Al(OiPr)3 species (Figure 1.5) [16]. Summing up these results it can be concluded that the isopropanolic solution of Al(OiPr)3 is dominated by the tetrameric Al(OiPr)3 species 1 (Figure 1.5) accompanied by a smaller amount of trimeric species.

1.2.2.2

Structural Changes at the Fluorination Process in Sols and Thin Gels

On the basis of the 27Al NMR spectra of isopropanolic Al(OiPr)3 solution, structural changes were followed after adding increasing amounts of HF/iPrOH solution to the aluminium isopropoxide solution according to Equation (1.4). In dependence on the molar Al:F ratio the appearance of the reaction mixture ranges from clear sols (4:1) to opaque gels (1:3). Figure 1.7 represents the 27Al and 19F NMR spectra obtained with rising content of fluorine (from a to d). It is obvious that in this order the amount of AlO6 species

Functionalized Inorganic Fluorides

12

(narrow line in Figure 1.4a,b; central units of 1 (Figure 1.5) is decreased. In the same manner a decreasing proportion of the sum over all fourfold Al species is found and, contrary to that, an increasing amount of fivefold coordinated aluminium species AlO5 (signals at about 35 ppm) [23].

27Al

a

19F

–156 ppm –147 ppm

a

–165 ppm –163 ppm

b

c

d

–161 ppm

b

–171 ppm

c

d

500 400 300 200 100 0 –100 –200 –300 –400 –100 (ppm)

–120

–140 –160 (ppm)

–180

–200

Figure 1.7 27Al NMR and 19F NMR spectra of different sols and wet gels(B0¼9.4 T). For all: solid line: experimental spectrum, dashed: simulation, dotted: decomposition. From a to d increasing content of fluorine. Molar ratios Al: F: (a) 4 : 1, (b) 2 : 1, (c) 1 : 1, (d) 1 : 2. (Reprinted with permission from [23] Copyright (2007) American Chemical Society.)

All 19F spectra are characterized by a group of three sharp signals (–161 ppm, –163 ppm, –165 ppm) with different intensities. All these signals are in a typical region for fluorine bounded on aluminium centres in a mixed oxygen-fluorine coordination with different fluorine ratios [13, 24–28]. With higher fluorine content, the intensity of the 19F NMR spectrum is more and more dominated by a broad peak at about –160 ppm (Figure 1.7, 19 F,d). These line-broadening effects result mainly from 19F-19F homonuclear dipolar couplings ending up in one broad peak in the static 19F NMR spectrum for the gel with molar ratio Al:F as 1:3. 1H and 13C NMR spectra of sols and gels show two main effects

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

13

with increasing fluorine supply: the portion of bridging isopropoxide groups is decreasing while the portion of terminal isopropoxide groups is affected only with higher fluorine supply [16, 23]. Obviously, the fluorination starts by protonation of bridging isopropoxide groups with the consequence of a line broadening of the corresponding signals, while the intensities of signals of terminal groups first remain constant. This first step is also supported by recently accomplished DFT calculations [23, 29]. A subsequent fluorination step may involve an attack of fluorine ions or HF on the central Al atoms of 1 (Figure 1.5) and the substitution of the protonated isopropoxide group by fluorine. 1.2.2.2.1 Isolated Single Crystals as Intermediates Evidence for a stepwise introduction of fluorine into the coordination sphere of Al is also given by the successful isolation of single crystals of partly fluorinated aluminium as well as magnesium alkoxide fluorides, Al3(OiPr)8FDMSO (Figure 1.8a) [9], Al3(OiPr)8FPy (Figure 1.8b) [30] and Mg6F2(OMe) 10(MeOH)14 (Figure 1.11) [31, 32]. In each case the isolation of single crystals works only if the fluorine supply is very low, i.e. Al:F or Mg:F > 1.

O(6) O(8)

O(1)

Al(3) O(7)

O(3)

O6 Al2

Al(2) O(5)

O(2)

O2

O(4) Al(1)

Al3

F

a

Al1

O4

O(9) O7

O5

O1

O3

O8

N1 F1

S

b

Figure 1.8 Crystal structure of (a) Al3(OiPr)8FDMSO (Reprinted with permission from [9] Copyright (2005) RSC); (b) Al3(OiPr)8FPy (Reprinted with permission from [30] Copyright (2008) Humboldt University of Berlin.)

Three distinguishable aluminium sites and one fluorine site are expected for Al3(OiPr)8FPy in the 27Al and 19F MAS NMR spectra, which are given in Figure 1.9. Simulation of the 27Al NMR spectrum supports this assumption and the decomposition obtained is given in Figure 1.9. The chemical shift values are close to those of crystalline Al(OiPr)3 emphasizing their structural similarity. As a consequence of the coordination of fluorine and a solvent molecule the originally higher symmetric AlO6 unit, now AlO4FPy, has considerably larger quadrupolar parameters [30]. The closeness to the Al(OiPr)3 structure may also be seen by comparing the two 1H-13C CP MAS NMR spectra (Figure 1.10).

14

Functionalized Inorganic Fluorides –160 ppm

100

0

–100

(ppm)

0

–20 –40 –60 –80 –100–120 –140 –160 –180 –200 –220 –240 –260 –280 –300 –320

(ppm)

Figure 1.9 27Al and 19F MAS NMR spectra of Al3(OiPr)8FPy; ( rot ¼ 25 kHz, B0 ¼ 9.4 T) (Reprinted with permission from [30] Copyright (2008) Humboldt University of Berlin.)

80

70

60

50

40

30

20

10

40

30

20

10

(ppm)

70

60

50 (ppm)

Figure 1.10 1H-13C CP MAS spectrum of Al3(OiPr)8FPy (bottom) in comparison with analogous spectrum of its precursor Al(OiPr)3 (top) ( rot ¼ 10kHz, B0 ¼ 9.4 T) (Reprinted with permission from [16] Copyright (2006) Humboldt University, Reprinted with permission from [30] Copyright (2008) Humboldt University of Berlin.)

Those CHO and CH3 groups located on bridged Al-O-Al positions (s. 1, 3 in Figure 1.5), i.e. at the positions > 65 ppm and < 25 ppm, respectively, are especially strongly affected by the new substituents on the central aluminium site. In contrast, the terminal groups are, as expected, less influenced.

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

15

Likewise, the polymeric nature of magnesium methylate results in the bulky Mg6F2(OMe)10(MeOH)14 (Figure 1.11), the synthesis of which can easily be reproduced. In this case the structure is also very similar to that of the fluorine-free compound Mg(OCH3)2 [33]. To preserve the cube-like shape, two fluorine ligands have to be introduced. The high symmetry of the crystal, however, results in one narrow 19F signal at 174 ppm, as shown in [32, Figure 5 therein].

012

011

010 08

07 04

Mg3 09

Mg2 03

06

05

F 01

Mg1

02

Figure 1.11 Crystal structure of Mg6F2(OMe)10(MeOH)14 (Reprinted with permission from [31] Copyright (2008) Wiley-VCH.)

1.2.2.3

Changes in Dry Gels with Progressive Fluorination

Isolated single crystals and their structures only give an idea of very early steps of fluorination, producible only with an understoichiometric fluorine supply. Changes in the local aluminium and fluorine coordination in solutions, sols and thin gels are

16

Functionalized Inorganic Fluorides

highlighted above (Figure 1.7). Drying of such thin and thick wet gels with different Al:F molar ratios leads exclusively to X-ray amorphous materials. However, based on the relevant liquid state NMR spectra (Figure 1.7), fundamental modifications can also be expected for the local coordinations in these solids. Together with the data already presented, a consistent mechanism of the fluorination process is then deducible. Taking all results into account, three stages of the sol-gel fluorination process can be identified, which are shown in the following. The first stage is primarily represented by samples with molar ratios Al to F higher than 1 (meaning low F-contents). Samples with molar ratios 1:1 and 2:3 (stage 2) mark for the solids the changeover to the third stage, the aluminium isopropoxide fluoride xerogels, samples with Al:F ratios lower than 0.5. 1 H-13C CP MAS, 27Al and 19F MAS spectra of dried gels with varying composition, as depicted in Figures 1.12 and 1.13, illustrate these findings. Beside a general shift of the 13C

Al O

Al(OiPr)3

Al

O

Al

O

CH

CH

Al CH3

Al : F

4:1

2:1

1:1

2:3

1:2

1:3

80

70

H

F O

CH

Al

60

H

50

40 δ13C /ppm

30 H

O

CH

20 H

O CH3

10 F O

Al CH3

Figure 1.12 1H-13C CP MAS NMR spectra (central transitions) of Al(OiPr)3 and aluminium isopropoxide fluoride solids prepared with different molar ratios Al: F as given in the figure ( rot ¼ 10 kHz, B0 ¼ 9.4 T) (Reprinted with permission from [34] Copyright (2009) American Chemical Society.)

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

17

signals to higher fields (CH3 groups) and lower fields (CHO groups), a considerable broadening effect is observable for all bridging CHO – and CH3 groups, and after that, with a higher proportion of fluorine, also a broadening and finally disappearing of the signals of the terminal groups (Figure 1.12) [34]. The spectral changes (peak maxima, line forms) deducible only from the 1H MAS NMR spectra are only minor and the superimposition of the several signals makes it difficult to discuss them separately [34]. 27Al and 19 F MAS NMR spectra however address local changes in the solids during continued fluorination very clearly. As mentioned above, the greatest change can be observed by introducing more and more fluorine and passing the molar ratio Al to F equal to one (see also Figure 1.13). Whereas the spectral features of the initial Al(OiPr)3 are still present in the 27Al MAS NMR spectra of samples with low fluorine content, a rising signal at 38 ppm is detected, which is provoked by a tetrahedrally coordinated aluminium site in the proximity to fluorine as evidenced by 19 27 F Al CP MAS NMR experiments [34]. Their 19F MAS NMR spectra (Figure 1.13) are dominated by very sharp signals (FWHM less than 1 kHz), which indicate ordered (‘crystal-like’) local structures. Most of the species in these solids are more-or-less

Al(OiPr)3

4:1

2:1

–162 –156 –148 –138

1:1

–171 –182

2:3

1:2

1:3 100

75

50

25

0 –25 δ27Al /ppm

–50

–75 –100 –50 –75 –100 –125 –150 –175 –200 –225 –250 δ19F /ppm

Figure 1.13 27Al and 19F MAS NMR spectra of Al(OiPr)3 and aluminium isopropoxide fluoride solids prepared with different molar ratios Al: F as given in the figure ( rot ¼ 25 kHz, B0 ¼ 9.4 T) (Reprinted with permission from [34] Copyright (2009) American Chemical Society.)

18

Functionalized Inorganic Fluorides

isolated; no proximity of the certain F-sites to each other can be stated from 19F-19F spin exchange experiments [34]. The 3QMAS NMR spectra of samples d and e (Figure 1.13) indicate the existence of a set of different AlFx(OiPr)4x – AlFx(OiPr)5x and AlFx(OiPr)6x – species (for the latter x ¼ 3–5) [34]. They are also responsible for the remarkable line-broadening effects in the corresponding fluorine spectra. The existence of certain fourfold and fivefold coordinated AlFx(OiPr)CN-x species as intermediate structures in aluminium isopropoxide fluorides was also unambiguously shown utilizing, for the first time, ultra high-field MAS NMR at magnetic fields B0 up to 21.1 T [35]. Moving to the third stage, a more and more stable network is formed with Al:F ratios equal to 1:2 and 1:3. The amount and spread of fourfold and fivefold coordinated Al-species decreases, ending up with sixfold AlFx(OiPr)6–x species (x ¼ 4 and 5) as deduced from the chemical shift correlation graphs [2628, 36, 37]. The comparison of the development of the intensities of single species with rising fluorination degree with the general development of the certain contributions of the appropriate 19F MAS NMR spectra allows a simple correlation of Al and corresponding F-species. Besides, a variety of possibly terminal fluorine-sites are evident for the highly disordered and amorphous aluminium isopropoxide fluorides in the up-field part of the spectra [35]. 1.2.2.4

Structure of Wet and Dry Aluminium Alkoxide Fluoride Gels AlF2.3OiPr0.7.xiPrOH – A Comparison

An attribution of local structures in wet gelatinous and air-sensitive fluoride gels implies many difficulties, which made the development and testing of inserts for MAS experiments necessary. Alternatively, low-temperature MAS NMR experiments at temperatures below the melting point of the solvent used allow the gel to be filled directly into the rotor [38]. 27Al and 19F MAS NMR spectra of a wet gel recorded at a spinning speed of 10 kHz both in a quartz insert and at low temperature are shown in Figure 1.14 together with the corresponding static spectra. A broad signal around 0 ppm was obtained for the 27Al static NMR spectrum (Figure 1.14a), completely covering the region for AlO4, AlO5, AlO6, AlOxFy or AlF6 species. The very broad static 19F signal is superimposed by narrow lines at –150 ppm and –171 ppm. The latter can be assigned to ‘free’ and mobile F- ions. Rotation at 10 kHz discloses a substantial line narrowing with a 27Al line at –16 ppm with a shoulder and an asymmetric decay in the high-field part. The 19F MAS NMR spectra of the wet gel observed using the quartz insert and applying low temperature are also comparable (Figure 1.14). A broad main signal is visible at –163 ppm together with spinning side bands superimposed by narrow peaks as contributions from the solvent. Comparing the MAS spectra of wet and dry gels, it becomes obvious that most of the structural features are already preformed in the jelly-like gel and in principle conserved and strengthened in the dry gel. This holds for all studied nuclei (19F, 1H, 13C, 27Al [38]). For 27Al this comparison is given in Figure 1.15 [38]. The shoulder and the position of the 27Al central lines as well as the wide spread of spinning side bands are characteristic features for both wet and dry gels. Sixfold coordinated Al-species (AlFxO6-x) are the dominating units in their structure. Second-order quadrupolar broadening is the main line-width factor of the Al signals. Four different structural units could be assigned by additional MQMAS experiments (Figure 1.16, [40]).

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

19

c

27Al

19F

a

–150 ppm –171 ppm

b

static, rotor

static, glass –16 ppm

*

**

10 kHz, 150 K 400

*

10 kHz, quartz * * *

1000 800 600

*

*

10 kHz, quartz

* *

10 kHz, 155 K * 200

0

–200 –400 –600 –800 –1000 100 50

0

*

*

–50 –100 –150 –200 –250 –300 –350 –400

(ppm)

(ppm)

Figure 1.14 MAS NMR spectra (B0 ¼ 9.4 T) of the wet gel using glass and quartz as insert materials in comparison with the respective static spectra (na: number of accumulations): a) 27Al: rot ¼ 10 kHz, quartz insert (na: 150000);  rot ¼ 10 kHz, frozen wet gel in the rotor at 150 K (na: 1000); (b) 19F:  rot ¼ 10 kHz, quartz insert (na: 192); frozen wet gel in the rotor at 155 K (na: 48);*: spinning side bands; (c) picture of a wet gel after rotation at 10 kHz in a quartz insert. (Reprinted with permission from [38] Copyright (2007) Elsevier Ltd.)

The reconstruction of the 27Al MAS NMR spectrum of the xerogel was possible with the four sites (A-D, Figure 1.16) obtained by analysis of the 3QMAS spectrum. [40]. These sites are attributed to AlF0-2O6-4 units (A), AlF4-5O2-1 units (38 %, B and C) and AlF5O1 units (53 % D) in the network [27, 40]. The existence of already immobilized –OiPr-groups, incorporated into or associated on the network, can be unambiguously proven for the wet gel by 13C CP MAS experiments [38]. In addition, the more rigid structure of the dry gel allows distinctions to be made between matrix groups and associated immobilized iPrOH molecules [38, 41]. The main structural features of the dried gel are already built up in the jelly-like gel and change only little in aging and drying processes. 1.2.2.5

Network Formation with Increasing Fluorine Content

Bearing in mind all NMR spectra recorded for sols, thin wet gels and dried gels (Figures 1.7, 1.12–1.16) in comparison with those of solid and liquid educts, the following mechanism can be derived for the fluorolytic sol-gel reaction. In the first step of the synthesis route of HS-AlF3, which is usually applied, the tetrameric structure of Al(OiPr)3

20

Functionalized Inorganic Fluorides

wet gel 160 K

a

100

75

50

25

0

–25 –50 (ppm)

–75 –100 –125 –150

Xerogel

b

νrot = 25 kHz

100

75

50

25

0

–25

–50

–75 –100 –125 –150

(ppm)

Figure 1.15 27Al MAS NMR spectra of the wet and dried gel (xerogel) with focus on the central region (400 MHz spectrometer; na: number of accumulations): (a) comparison of 27Al MAS NMR spectra of the frozen wet gel with (solid line, 160 K) and without (dotted line, 155 K) 19F (cw) decoupling at cryo-temperatures, for both:  rot ¼ 10 kHz, na: 1000; (b) possible different contributions to the spectrum of the xerogel including the n ¼ 0 spinning sideband of the satellite transition: (- - -). (s. also Neuville, Massiot [39]); experimental spectrum (——); simulated spectrum () and decomposition ( -  -  -) (see also Table 2 in [38]). (Reprinted with permission from [38] Copyright (2007) Elsevier Ltd.)

as dominating species (Figure 1.5, 1) represents the starting point for the further reaction pathway and finally for the formation of the gel network. The fluorination begins with a protonation of bridging isopropoxide groups, which makes a subsequent attack of fluorine ions or HF to central Al atoms easy. As result, a substitution of protonated isopropoxide groups by fluorine occurs. At early fluorination states (state 1, Scheme 1.4) unconverted Al(OiPr)3-molecules exist in the sols and gels to some degree – either in its tetrameric or in its linear trimeric form. In the following the latter may be partly fluorinated and, if stabilizing donors (D) are

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

δMQMAS(27Al) in ppm

–40 D –20

A

qc

21

cs C

0 20 B 40 20

0

–20 –40 δMAS(27Al) in ppm

–60

Figure 1.16 3QMAS spectrum of the aluminium alkoxide fluoride gel (Al:F ¼ 1:3) acquired at 14.1 T (Reprinted with permission from [40] Copyright (2009) American Chemical Society.)

accessible, even crystals of Al3(OiPr)8F•D can be isolated (Scheme 1.4, 3). Nevertheless, for aluminium isopropoxide fluorides resulting from sols without donor molecules the formation of ordered similar species Al3(OiPr)8F•iPrOH is presumable as evidenced, for example, by the corresponding sharp signals centred at –160 ppm and –171 ppm (19F). The early formation of tetrahedrally coordinated species AlF4 can be demonstrated for solids prepared with low fluorine supply. This species does not seem to exist in the corresponding sols. Instead, it is plausible that these species are stabilized as solvated species AlFx(HOiPr)6x3x (x ¼ 4) in the sols. Drying results in a partial loss of the solvent molecules and the formation of ‘incorporated’ AlF4-species. Additionally, further linking processes of the AlF4(HOiPr)2 species lead to the formation of bigger units (beginning of a gelous network), which, as a consequence, are predominantly built of AlFx((HOiPr)6-xoctahedra (x ¼ 3 to 5). The rise of the F-content (molar ratio Al: F < 1) leads to an irregularly strongly distorted and disordered solid. The corresponding sol consists of a loose network, imaginable in solution as presented earlier in [23]. Vacuum drying leads to cleavage of iPrOH molecules coordinating to Al-species existing in these sols and gels, forming a variety of fourfold, fivefold and sixfold coordinated Al-species like AlFx(H)OiPr)4x, AlFx((H)OiPr)5x and AlFx((H)OiPr)6x as identified by 3QMAS and ultra high-field MAS NMR measurements [34, 35]. The units are sterically separated; as in the 19F-19F EXSY NMR experiments, nearly no spin exchange could be observed ([34], state 2, s. Scheme 1.5). Finally this gelous network is strengthened and stabilized by cross-linking with higher fluorine supply. Vacuum drying leads then to xerogels that consist predominantly of sixfold coordinated AlFx((H)OiPr)6x species (x: 35)[38, 40]. Coordinating solvent molecules are stabilized by the formation of H-bridges. The local structures of the xerogel are preformed in the wet gel. Nevertheless, ordered local structures are also observable for the xerogels. The derived mechanism is schematically depicted in Schemes 1.4 and 1.5.

22

Functionalized Inorganic Fluorides O

State 1

O side reaction

O

O Al O

O

Al

if donor present isolable

O Al O F D

Al O

3

species present in isopropoxide fluoride sols with a ratio F: Al lower 1… O

Al O O

O O

O

O O Al O O O Al O O 1

Al O

O

Al

O*

F

O

O O Al O O

Al

F

F F

*O

O*

F

F F *O Al

2

The loss of solvent leads to formation of AlF4 species along with un converted Al(OiPr)3 1 and cross-linked AlFx(OiPr)6–x species.

F

t

F

Al

F F

F

Al

O F O Al F * F F

F F

F

F O* Al F O F F Al O O Al O* F F F * F F

F

Al

F

F

F

F

– O*

O

…and in the corresponding solids.

F

O Al

F

Scheme 1.4 Derived mechanism for the beginning sol-gel fluorination process with low fluorine supply (O*: iPrOH; O: -O-iPr). (Reproduced from [34] by permission of the American Chemical Society.) State 2

wet sol:

F

F *O

O Al

F

solid:

F

F

F F F O F Al *O Al O F F F * Al F F F F F F Al F O F F O* Al F O F F Al O O Al O* F F F * F F

and further

– O*

O Al

O

F F

Al

F

F F F F F F F O F Al Al F *O Al F O F *O Al F F * Al F F F F F F F F F Al F O F F Al F F O O* Al F F F Al Al O F O F O* Al F F * F F F F Al O O Al F F F * F F *O

irregular solid

O

minor ? O Al O

Al O O O O Al O O* F 3` (D = O*)

indications for formationof a similar compound like 3

Scheme 1.5 Derived mechanism for a progressive fluorination degree (reproduced from [34] by permission of the American Chemical Society)

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

1.2.2.6

23

Structural Characterization of Sol-Gel Derived High-surface Fluorides

1.2.2.6.1 Vibrational Spectroscopy of HS-AlF3 The IR transmission spectra of HS-AlF3 were recorded and compared with those of aluminium isopropylate and of dry aluminium fluoride gel – i.e. the Al moieties involved in HS-AlF3 synthesis (Figure 1.17) [8], as well as of crystalline AlF3 phases (Figures 1.18 and 1.19) [42]. In Figure 1.17 the CH3 and CH as well OH absorption bands of isopropoxide and/or isopropanol can be seen at wave numbers higher than 2800 cm-1 in the spectra of aluminium isopropylate and of dry aluminium fluoride gel, whereas the broad band in the HS-AlF3 spectrum can be attributed (i) to water adsorbed in course of sample preparation due to the high Lewis acidity, and (ii) to water residues present in the bulk as a consequence of the formation of diisopropylether at postfluorination. At low wave numbers there are the absorption bands of Al-O and Al-F valence vibrations (360–700 cm-1) and of C-C frame vibrations of the organic constituents. These spectra show the disappearance of organic components converting the dry gel, the ‘precursor’, into HS-AlF3.

Transmission

HS-AlF3

precursor

Al(Oi Pr)3

4000

3500

3000

2500

2000

Wave number

1500

1000

500

(cm–1)

Figure 1.17 IR spectra of Al(OiPr)3, the AlF2.3OiPr0.7.xiPrOH gel, and HS-AlF3 (Reprinted with permission from [8] Copyright (2007) Springer Science þ Business Media.)

The amorphous, highly disordered state of HS-AlF3 follows convincingly from a comparison of its IR spectra with those of -AlF3 shown in Figure 1.18. Since crystalline -AlF3 also gives a well resolved Raman spectrum (Figure 1.19) all the distorted X-ray amorphous phases of HS-AlF3 do not yield any useful information. The very broad peaks in the IR-spectrum of HS-AlF3 are obviously the consequence of its amorphous state and can be interpreted as superposition of many unresolved peaks, also covering the range of vibration bands of the - and -AlF3 phases. 1.2.2.6.2 X-ray Diffraction and TEM Investigations of HS-AlF3 In contrast to vibrational spectroscopy, the X-ray diffraction patterns of HS-AlF3 and of its precursor, the AlF2.3OiPr0.7.xiPrOH dry gel, show no peak and therefore only indicate the

24

Functionalized Inorganic Fluorides

HS-AlF3

Transmission

β-AlF3

α-AlF3

0

200

400 600 800 Wavenumber/cm–1

1000

1200

Figure 1.18 IR-transmission spectra of a-AlF3, -AlF3, and HS-AlF3 (Reprinted with permission from [42] Copyright (2007) American Chemical Society.) 157

30 000

20 000

10 000 5000

478

383

15 000

96

Intensity [arb.units]

25 000

100

x25

200

300

400 500 Wavenumber [cm–1]

600

700

800

Figure 1.19 Raman spectrum of a-AlF3 (rhombohedral phase) (Reprinted with permission from [42] Copyright (2007) American Chemical Society.)

X-ray amorphous state of these samples. Only after heating above the crystallization temperature of about 836 K do peaks corresponding to -AlF3 become apparent (Figure 1.20). In the TEM micrograph (Figure 1.21) both the agglomeration of nano-particles to microparticles as well as the partial crystalinity of the nano-particles can be seen. 1.2.2.6.3 Solid State NMR of HS-AlF3 Subsequent drying of the xerogel in vacuum results in a noticeable loss of –OiPr-groups and associated iPrOH molecules, which can easily be followed by 1H-13C CP MAS

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

25

Intensity (au)

(a)-precursor (b)-precursor in N2 at 350 °C (c)-HS-AlF3 (d)-precursor in N2 at 700 °C (α-AlF3)

(d) (c) (b) (a)

10

20

30

40

50

60

2θ

Figure 1.20 X-ray powder difractograms of (a) AlF2.3OiPr0.7.xiPrOH dry gel, (b) AlF2.3OiPr0.7.xiPrOH dry gel heated to 623K, (c) HS-AlF3, and (d) AlF2.3OiPr0.7.xiPrOH dry gel heated to 973 K (Reprinted with permission from [8] Copyright (2007) Springer Science þ Business Media.)

Figure 1.21 TEM micrograph of HS-AlF3 (Reprinted with permission from [6] Copyright (2003) Wiley-VCH.)

experiments [44]. Both 19F and 27Al MAS NMR spectra exhibit after removal of the organic components the typical shape as recorded for HS-AlF3. The maximum broad 19F MAS NMR signal lies typically at –165 ppm [41]. The effect of such a drying process is shown exemplarily for 27Al in Figure 1.22. The spectrum on the bottom (Figure 1.22) is almost identical with the central line usually measured for HS-AlF3. In contrast to the xerogel (AlF2.3OiPr0.7.xiPrOH) the 27Al MAS NMR signal line widths do not narrow

26

Functionalized Inorganic Fluorides

much upon increase of the applied magnetic field. Therefore the comparison of MQMAS experiments at different magnetic field strength is most important for an assignment of various species. Although the data analysis is very difficult, the presence of various signals was revealed [40]. A 3QMAS spectrum of HS-AlF3 is given in Figure 1.23.

70 °C, 2 h

100 °C, 2 h -–15 ppm

150 °C, 2 h

170 °C, 4 h 100

50

0

–50

–100

–150

(ppm)

Figure 1.22 27Al MAS NMR spectra (central lines) of a AlF2.3OiPr0.7.xiPrOH gel thermally dried in a vacuum at different temperatures ( rot ¼ 25 kHz, B0 ¼ 9,4 T) (Reprinted with permission from [44] Copyright (2009) Humboldt University of Berlin.) –40 qc δMQMAS(27Al) in ppm

B'

cs

–20 A' 0 20 40 20

0

–40 –20 δMAS(27Al) in ppm

–60

Figure 1.23 3QMAS NMR spectrum of HS-AlF3 acquired at 19.9 T (Reprinted with permission from [40] Copyright (2009) American Chemical Society.)

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

27

The line shape analysis of the 27Al MAS NMR spectrum of HS-AlF3 required more than the two Al-sites extracted from the 3QMAS spectrum (Figure 1.23). A successful fit was only possible by implementation of two further Al-sites. For all of them the simulation required the Czjzek-model, i.e. distributions of quadrupolar parameters were used [40]. Now 65 % of the integral intensity of the Al signal belongs to AlF5-6O1-0 and AlF6 units, respectively. Residues of AlF4-5O2-1 units (25 %) and AlF0-2O6-4 units (11 %) are still present [40].

1.2.3

Exploring Properties

The most remarkable property of HS-AlF3 is its outstanding Lewis acidity [6, 8, 9, 45], which is far higher than that of AlCl3, as will be shown later. It is interesting to note that on a theoretical basis for an isolated ‘AlF3-molecule’ already a Lewis acidity was predicted ranking among the highest ones at all [46]. In the following a variety of investigations showing the very high Lewis acidity of HS-AlF3 will be presented and discussed.

1.2.3.1

Adsorption of Probe Molecules

The strength and/or nature and/or amount of acid sites accessible at the surface of a solid can principally be determined measuring the interaction with a basic probe molecule. The higher the acidity of the solid the lower can be the basicity of the probe and vice versa. For investigations of HS-AlF3, pyridine and its derivatives, ammonia and carbon monoxide, were employed. Their interactions with the solid have been analysed studying in case of ammonia desorption with increasing temperature, Temperature Programmed Desorption TPD, and in case of the other probes via IR spectroscopy. NH3-TPD: HS-AlF3 was saturated with gaseous NH3 followed by flushing with N2. Upon gradually heating the sample up to 773 K, i.e. below the crystallization temperature, the adsorbed NH3 is gradually released as shown in Figure 1.24. Compared to the well-known solid Lewis acid -AlF3 the desorption from HS-AlF3 occurs up to much higher temperatures corresponding to a much higher Lewis acidity, and the total amount of NH3 is also much higher, indicating a higher number of acidic sites per gram. Pyridine adsorption: The chemical nature of the acid sites can be seen from photoacoustic infrared spectra (PAS) of adsorbed pyridine (Figure 1.25). Frequencies and intensities of the IR bands show for HS-AlF3 (almost) only Lewis acid sites, whereas a HS-AlF3 sample showed after treatment with HF predominantly Brønsted acid sites (Figure 1.25b). Carbon monoxide adsorption: Carbon monoxide behaves towards a strong acid as a weak base the interaction of which can be investigated monitoring the CO stretching region of the IR absorption spectrum of the absorbed CO. The stronger the acid the more is the CO IR frequency blue-shifted. HS-AlF3 shows the strongest blue shift ever reported for a solid acid (for details see Chapter 3) indicating that it is, next to ACF (aluminium chlorofluoride), the strongest solid acid of all [47].

28

Functionalized Inorganic Fluorides HS-AIF3 beta-AIF3

35 000 30 000 25 000 20 000 15 000 10 000 5000 0 –5000

0

100

200

300

400

500

Temperature/°C

Figure 1.24 NH3-TPD of HS-AlF3 in comparison to -AlF3 (Reprinted with permission from [7] Copyright (2007) Royal Society of Chemistry.)

Figure 1.25 PAS of pyridine adsorbed on (a) HS-AlF3, (b) HS-AlF3 exposed to HF, (c) HS-AlF3 prepared by post-fluorination with HF, and (d)sample as in(c) additionally heated in CHClF2/N2 flow; peaks at (B) are indicative for Brønsted acidity, and peaks at (L) for Lewis acidity (Reprinted with permission from [8], Figure, Copyright (2007) Springer Science þ Business Media.)

1.2.3.2

Catalytic Test Reactions

Reactions which have to be catalysed by a Lewis acid to proceed can be used as testreaction for assessment of the acidity of a material under study. Equations (1.5) to (1.8) show four reactions, the use of which has been reported [8, 45]: 5 CCl2 F2 ! CCl3 F þ 3 CCl3 F þ CCl4

(1:5)

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

29

3 CHClF2 ! CHCl3 þ 2 CHF3

(1:6)

CBrF2 CBrFCF3 ! CF3 CBr2 CF3

(1:7)

CCl2 FCClF2 ! CCl3 CF3

(1:8)

The educts of the reactions in Equations (1.5) and (1.6) were used as postfluorination agents in the course of high surface metal fluoride preparation, especially that of HS-AlF3. As soon as first parts of the dry Al-F-gel are converted into HS-AlF3, a dismutation reaction according to Equations (1.5) or (1.6) starts, which can be detected by GC. Thus, these fluorinating agents act conveniently as detectors for Lewis acidity. However, the reactions in Equations (1.5) and (1.6) already proceed in the presence of comparable weak Lewis acids; they are therefore no measure of high acidity. The isomerization of CBrF2CBrFCF3 (Equation 1.7), on the other hand, proceeds only under the catalytic action of the strongest known Lewis acids, i.e. antimony pentafluoride or aluminium chlorofluoride, and is also catalysed by HS-AlF3 but not by AlCl3 [8, 45]. Thus, the isomerization reaction gives convincing evidence for the exceptional high Lewis acidity of HS-AlF3. As AlCl3 is widely used in organic synthesis as a Lewis acidic catalyst a further comparison of its Lewis acidity with that of HS-AlF3 was of interest. Studying the isomerization of CCl2FCClF2 (Equation 1.8) as test reaction, AlCl3, HS-AlF3, ACF, -AlF3 and -AlF3 have been compared regarding their respective catalytic activity. It was found that ACF and HS-AlF3 were catalytically very active, and -AlF3 and -AlF3 not at all, as expected. Surprisingly, AlCl3 was also not primarily active but became active only under conditions and after some time as was needed to convert AlCl3 into ACF. The experimental result was interpreted based on theoretical investigations assuming that under-coordinated Al atoms, which are a result of the high degree of disorder, are responsible for the Lewis acidity of HS-AlF3 [45]. Radiotracer investigations: Radiotracer experiments, which also gave evidence for the exceptional Lewis acidity of HS-AlF3, are discussed in Chapter 3.

1.2.4 1.2.4.1

Possible Fields of Application Range of Metal Fluorides Obtainable via Sol-Gel Fluorination

The fluorolytic sol-gel synthesis of metal fluorides was originally developed and explored for aluminium fluoride, which was a piece of luck since both the stepwise synthesis and the properties of HS-AlF3 showed the influence of the new synthesis process. Thus, almost immediately after exploration of HS-AlF3 other binary metal fluorides have been similarly prepared and attempted syntheses of more complex systems started. Binary metal fluorides: Many binary metal fluorides have been prepared via sol-gel fluorolysis [4, 7]. The applicability of the synthesis process is primarily limited by the ready availability of soluble metal alkoxides. However, the synthesis of metal fluorides, of which the metal ions are very weak Lewis acids, typically does not result in sol formation but finely dispersed solid fluorides, the XRD of which reflect the aimed-for compounds.

30

Functionalized Inorganic Fluorides

Thus, upon fluorolysis of the tert-butoxides of Li, Na, K and Cs in THF solution, only with LiOtBu and NaOtBu did gel formation occur whereas with K- and Cs- alkoxide immediate precipitation was observed. The dried products of all these alkali metal ions reflected the corresponding fluorides in XRD and also gave some indications of the respective hydrogenfluorides due to their higher thermodynamic stability [48]. The most thoroughly investigated example besides HS-AlF3 is HS-MgF2 [11]. Other high surface area alkali earth fluorides prepared are HS-CaF2 and HS-BaF2 [49]. Mixed metal fluorides: Prompted by a hypothetical model by Tanabe explaining the Lewis acidity of guest-host metal oxide systems [50] and its adaptation for fluoride systems [51], guest-host mixed metal fluoride systems with HS-MgF2 as host have been prepared (see also Chapter 3). If the radii are comparable, the metal ions guest ions will probably occupy places of Mg2þ ions. If the guest ion has a higher positive charge than Mg2þ, Lewis acidity should be created in accordance with the model. This way, solid potential metal fluoride catalysts with tunable Lewis acidity are accessible. Thus, with up to 20 mol % of Fe3þ, Ga3þ, V3þ, In3þ, and Cr3þ as guest components, solid solutions with HS-MgF2 as host have been prepared [52–54]. The synthesis of such systems basically followed the fluorolytic sol-gel route described above, however, in some experiments compounds other than alkoxides have been employed as guest components to be added to the Mg alkoxide solution. Analyses of these mixed systems by XRD, 19F MAS NMR and photoelectron spectroscopy showed no evidence of the guest compounds but gave only of the typical spectral data of MgF2, which however, were not identical with those of crystalline MgF2. Obviously the guest metal ions were incorporated into the MgF2 lattice as expected. ESCA investigations in the Cr3þ/MgF2 system (Figure 1.26) showed, for both Mg 1s and F 1s, binding energies near those of pure MgF2 [54]. A shift that is only small to lower binding energies gives evidence that the chemical environment of Mg and F in the mixed systems is influenced by Cr3þ. Complex metal fluorides: Complex metal fluorides are especially interesting compounds because of their physical properties, which are suitable for applications such as lasers. Conventional syntheses typically employ thermal methods starting from the respective metal fluorides. Complex metal fluorides can also be conveniently prepared via sol-gel fluorination. The sol-gel syntheses start similar to that of host-guest systems from mixtures of the respective metal alkoxides but in the respective stoichiometric ratio, which are subjected to fluorolysis. Contrary to the host-guest systems, where the host lattice is preserved, complex systems, i.e. fluorometallates, have their specific crystal structure different from those of the respective binary metal fluorides they are composed of. Examples of complex metal fluorides prepared via sol-gel fluorolysis are Li3AlF6, Na3AlF6, K3AlF6, KAlF4, CsAlF4, LiNa2AlF6 [55] and BaAlF5, K2MgF4 and BaMgF4 [49]. Since Al in AlF6 units is ideally shielded, such compounds do not show any Lewis acidity and are chemically very stable making them suited especially for nonchemical applications such as protective coating (see below). 1.2.4.2

Application Consequences of the Sol-Gel Synthesis

Specific chemical and physical properties of intermediate states of the sol-gel fluorolysis can be utilized for quite different modifications and applications. Upon reaction of metal

31

655

1339

1179

2482

654

1338

1178

2481

653

1337 MgF2 CrClMg15 CrOMg8 CrOMg15 CrOMg25 CrOMg50 crAcMg15 CrAcMg40

652 651 650 689

688

687 686 F ls (eV)

685

1336 1335 1334 684

Mg KLL (eV)

F KLL (eV)

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

2480

1177 MgF2 CrClMg15 CrOMg8 CrOMg15 CrOMg25 CrOMg50 crAcMg15 CrAcMg40

1176 1175 1174 1308

1307

1305 1306 Mg ls (eV)

1304

2479 2478

2477 1303

Figure 1.26 Chemical state plots for F 1s and Mg 1s orbitals of chromium(III)-doped MgF2 prepared from CrO3 (CrO), CrCl3 (CrCl) or Cr-acetat (CrAc) in different concentrations (Reprinted with permission from [54] Copyright (2005) Elsevier Ltd.)

alkoxides with understoichiometric amounts of hydrogen fluoride a defined part of the alkoxide groups remain and may be used for chemical modifications of the high surface metal fluorides. Such modifications include partial substitution of F by OH, chemical immobilization of metal oxide oxidation catalyst onto the metal fluoride and binding organic groups to the inorganic metal fluoride. Introduction of highly dispersed noble metals on metal fluoride is also possible. The sol state, i.e. the colloidal solution of a metal fluoride in a nonaqueous solvent, opens up the possibility of many interesting applications. Most important is the ability to prepare different coatings for very different applications. Thus, catalytic active metal fluorides can be conveniently deposited on supports, or thin metal fluoride layers, can be easily prepared for, e.g., optical or mechanical applications. Nano-sized metal fluorides: An optically clear metal fluoride sol contains particles the diameter of which is in the range of the wavelength of visible light. Therefore, nano-sized metal fluoride particles can be obtained from such sols, which are normally agglomerated, after removal of the solvent at moderate temperatures. The size of the individual particles has been shown by TEM and was derived from their XRD pattern to be below 5 nm [6, 49]. Such tiny particles represent a higher state of energy compared to more bulky material. As a consequence these materials can be easily pressed at room temperature to dense glasses. Thus, transparent glasses have been obtained from MgF2, CaF2, BaMgF4 [49] and Rb2NaAlF6 [56] employing pressures up to 1.1 GPa at room temperature. Nano-sized metal fluorides have a promising perspective as inorganic component of organic polymers. Due to their small size the particles are invisible even in transparent polymers but may, in possibly decisive ways, change physical properties such as dielectric constant, index of refraction or mechanical properties. However, preventing agglomeration of these nano particles is a challenging task, but one that is not easy to accomplish and that needs to be further developed. Oxide fluorides: The amount of alkoxide groups remaining after partial fluorination can be tuned over a wide range. These OR groups can subsequently be hydrolysed, i.e. substituted by OH, or thermally split off, also resulting in OH and, due to condensation reaction, in the formation of oxide groups homogeneously distributed within the fluoride matrix. The metal fluorides modified this way exhibit not only Lewis acidity but also some Brønsted acidity or

32

Functionalized Inorganic Fluorides

even basicity. Such bifunctional materials can be valuable solid catalysts (see below). This synthesis principle has been realized for magnesium oxide/hydroxide fluorides within a broad range of compositions [57]. It was found that preparations with low fluorine content, with nominal composition Mg(OH)1.2F0.8 and Mg(OH)0.8F1.2, were almost X-ray amorphous even after calcinations at 623 K whereas with higher F content the patterns of MgF2 appeared. In 19F MAS NMR there was a similar trend to be seen, the low F materials gave broad, complex signals, indicating the presence of many different fluorine species, which became more and more sharp and like those in MgF2 with increasing F content. The higher electronegativity of F compared to OH resulted with increasing F content in a reduced electron density at Mg indicated in increasing Mg 1s binding energy as seen in XPS analysis. Magnesium fluoride-based bifunctional materials have been successfully employed as heterogeneous catalysts for quite different reactions (see below) [57–59]. Metal oxides linked to metal fluorides: The performance of a solid catalyst depends to a substantial degree on its surface area, which can often be increased by suitable support, and then also on the chemical and texture properties of the support. Vanadium oxide-based catalysts are very useful in oxidation reactions because of the easy change between different oxidation states. In case of the technically important partial oxidation of organic compounds, a limited oxidation activity of the catalyst is needed to prevent total oxidation to CO / CO2 and H2O. As a consequence the organic educt has to be activated by, e.g., protonation due to Brønsted acidity of a catalyst support. By using a Lewis acidic support a more gentle activation of the educts should be possible resulting in less or no total oxidation. The sol-gel synthesis provides exquisite conditions for a very evenly distributed and highly dispersed VOx based catalyst deposited on Lewis acidic metal fluoride supports. Employing the principle of under-stoichiometric fluorination there will be OR groups or, after hydrolysis, OH groups available for chemical anchoring of the VOx species. This way aluminium fluoride-supported VOx catalysts with outstanding performance have been prepared and thoroughly characterized [60]. Noble_metals supported by high surface area metal fluorides: Noble metals especially platinum and palladium are useful catalysts for many reactions such as, e.g., hydrogenations. They are used in a highly dispersed state deposited on suitable supports. For use in hydrodehalogenation reactions the support has to be stable against attack by the hydrogen halogenide formed in course of the reaction. That requirement can be met using metal fluorides as support. MgF2 and AlF3 were often proposed [61]. For the preparation of very tiny noble metal particles supported on a metal fluoride, which has a large surface, the sol-gel fluorination provides excellent conditions. Starting from the metal alkoxide solution or from the already formed metal fluoride sol an organic solution of a suitable noble metal compound, such as the acetyl acetonate, is added and the mixture is worked up as normal for high surface metal fluoride preparation. Only an additional reduction step is necessary to obtain the catalyst. For a Pd0/CaF2 catalyst prepared accordingly the high Pd dispersion is shown in Figure 1.27. Organically modified metal fluorides: Silicon oxide and also metal oxide based inorganicorganic hybrid materials have received broad interest academically and also for technical applications [62]. The development of such hybrids aims to combine useful properties of the inorganic part, mostly the basis of the hybrid, with those of the organic part, whereby the two parts are chemically bound to one another. Basically, two types of synthesis routes

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

33

Figure 1.27 TEM micrograph of a Pd0/CaF2 catalyst showing Pd particles (black spots) of £10nm diameter (Reprinted with permission from [61] Copyright (2008) Royal Society of Chemistry.)

are used. Both can be adapted to metal fluoride systems. One route starts from organically modified molecular alkoxidic precursors, which are subjected to sol-gel hydrolysis; the other starts from preformed nanoscopic or macroscopic particles to which an organic moiety becomes linked. To transpose these synthesis principles to metal fluoride systems, the sol-gel fluorolysis can be conveniently employed attaching either an organic moiety via a HF stable linkage to metal alkoxides, which are consequently subjected to sol-gel fluorolysis. Another route comprises at first formation of colloidal nanoscopic particles by preparing a sol with under-stoichiometric amounts of HF, to which then organic moieties can be linked. First experiments with phenylphosphonate modified aluminium alkoxide have proved successful [30, 63]. Coatings: The sol state as primary result of the sol-gel fluorination process is very useful, because it consists of nano-sized, colloidal metal fluoride particles homogeneously dispersed in a nonaqueous liquid. Simply by applying the sol on any solid material wettable by the solvent and subsequent drying a metal fluoride layer is on the surface of the material obtained. Depending on the required properties and intended purposes of the layer and on the texture and geometry of the solid material, different methods can be used for the sol application such as, e.g., soaking, dipping, spin coating etc. Catalysis: For technical reasons it can be necessary to have a catalytic active solid material supported to improve, e.g., its mechanical stability and reduce its flow resistance when used in a flow reactor. The sol-gel fluorination synthesis provides a convenient way for depositing high surface area metal fluorides on supports. For example, HS-AlF3, which as fine powder makes problems when used as catalyst in flow systems, could be supported by g-Al2O3 whereby its Lewis acidity and consequently its catalytic activity remains almost unchanged [64]. For other catalytic applications, like micro-reactor techniques, deposition of catalytically active thin layers of metal fluorides is also of interest.

34

Functionalized Inorganic Fluorides

Optics: Probably the most important field of application for thin metal fluoride layers is in physics, especially in optics. A reduction of the reflexion of light at the surface of glass, well known for optical devices such as lenses for glasses and cameras, is even more important for solar energy utilization. Antireflective systems are formed of alternating layers of transparent low and high refractive index materials. However, even with a single layer, a very efficient antireflective system is possible in principle. For a wavelength g and an antireflective layer of thickness g/4 the reflexions at the air/layer surface and at the layer/glass interface have a phase-shift to each other of g/2, that is the precondition for extinction. Total extinction is only possible when the two reflexions are of the same intensity. This can be reached when the index of refraction of the antireflective layer is equal to the geometric mean of air (n ¼ 1.0) and glass (n ¼ 1.5), i.e. for n ¼ 1.225. Typically, the refractive indices of metal oxides are even higher than that of glass, whereas some metal fluorides such as MgF2 (n500 ¼ 1.38), AlF3 (n500 ¼ 1.35) and Na3AlF6 (n500 ¼ 1.33) have distinctly lower indices of refraction, although not as low as 1.225, together with an excellent transparency within a broad range of wavelengths. Consequently, MgF2 thin layers have already found much interest [65]. Methanolic MgF2 sols prepared as described above have been used for spin-coating of planar surfaces [66]. AFM investigations revealed that the layers obtained after drying consist of densely packed particles of 10 to 20 nm diameter as shown in Figure 1.28. The single or multiple MgF2 thin layers showed excellent homogeneity concerning thickness and index of refraction over the experimental range of 5 cm [67].

Figure 1.28 AFM image of a 3-fold deposited MgF2 layer on a silicon wafer, after calcinations at 300 C (area 1 x 1mm2). (Reprinted with permission from [67] Copyright (2008) Wiley-VCH.)

Protective coating: The facileness of preparing metal fluoride layers from nonaqueous metal fluoride sols also reduces the threshold of their use for protective coating. Homogeneous, mechanically stable metal fluoride layers can protect against UV radiation,

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

35

chemical and also mechanical impact, and can form a barrier against microbial attack. A CaF2 addition to lithium grease has already proved useful in the lab for friction reduction and extreme pressure properties making CaF2 layers, likewise, interesting. In conclusion, the new access towards nanoscopic metal fluorides via this recently developed fluorolytic sol-gel synthesis route opens a wide range of applications for metal fluorides due to the distinctive different properties of these nano materials.

References [1] W. Sto¨ber, A. Fink and E. Bohn, Controlled growth of monodisperse silica spheres in the micron size range, J. Colloid Interface Sci., 26, 62–69 (1968). [2] C. Brinker and G. Scherer, Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing, Academic Press, Boston, 1990. S. Sakka (Ed.), Handbook of Sol-Gel Science and Technology, Kluwer, Boston, 2005. [3] M. Dejneka, E. Snitzer and R. E. Riman, Spectroscopic characterization of Eu3þ-doped inorganic and alkoxide sol-gel derived fluorozirconate glass and zirconium fluoride gels, J. Non-Cryst. Solids, 202, 23–34 (1996). [4] St. Ru¨diger, U. Groß and E. Kemnitz, Non-aqueous sol-gel synthesis of nano-structured metal fluorides, J. Fluorine Chem., 128, 353–368 (2007). [5] S. Fujihara, T. Kato and T. Kimura, Sol-Gel processing and luminescent properties of rare-earth oxyfluoride materials, J. Sol-Gel Science Techn., 26, 953–956 (2003). [6] E. Kemnitz, U. Groß, St. Ru¨diger and S. C. Shekar, Amorphous metal fluorides with extraordinary high surface areas, Angew. Chem. Int. Ed., 42, 4251–4254 (2003). [7] St. Ru¨diger and E. Kemnitz, The fluorolytic sol-gel route to metal fluorides – a versatile process opening a variety of application fields, Dalton Trans., 1117–1127 (2008). [8] St. Ru¨diger, G. Eltanany, U. Groß and E. Kemnitz, Real sol-gel synthesis of catalytically active aluminium fluoride, J. Sol-Gel Sci. Techn., 41, 299–311 (2007). [9] St. Ru¨diger, U. Groß, M. Feist, H. A. Prescott, S. Chandra Shekar, S. I. Troyanov and E. Kemnitz, Non-aqueous synthesis of high surface area aluminium fluoride – a mechanistic investigation, J. Mater. Chem., 15, 588–597 (2005). [10] J. K. Murthy, U. Groß, St. Ru¨diger, E. Kemnitz and J. M. Winfield, Sol-gel-fluorination synthesis of amorphous magnesium fluoride, J. Solid State Chem., 179, 739–746 (2006). [11] S. Wuttke, G. Scholz, St. Ru¨diger and E. Kemnitz, Variation of sol-gel synthesis parameters and their consequence for the surface area and structure of magnesium fluoride, J. Mater. Chem., 17, 4980–4988 (2007). [12] D. Bradley, R. C. Mehrotra, I. Rothwell and A. Singh, Alkoxo and Aryloxo Derivatives of Metals, Academic Press, Boston, 2001. N. Ya. Turova, E. P. Turevskaya, V. G. Kessler and M. I. Yanovskaya, The Chemistry of Metal Alkoxides, Kluwer, Boston, 2002. [13] E. Kemnitz, U. Groß, St. Ru¨diger, G. Scholz, D. Heidemann, S. I. Troyanov, I. V. Morosov and M.-H. Leme´e-Cailleau, Comparative structural investigation of aluminium fluoride solvates, Solid State Sci., 8, 1443–1452 (2006). [14] N. Y. Turova, V. A. Kozunov, A. I. Yanovskii, N. G. Bokii and Y. D. Struchkov, B. L. Tarnopolskii, Physico-chemical and structural investigation of aluminium isopropoxide, J. Inorg. Nucl. Chem., 41, 5–11 (1979). [15] K. Folting, W. E. Streib, K. G. Caulton, O. Poncelet and L. G. Hubertpfalzgraf, Characterization of aluminium isopropoxide and aluminosiloxanes, Polyhedron, 10, 1639–1646 (1991). [16] R. Ko¨nig, Magnetresonanzuntersuchungen zur Ausbildung lokaler Strukturen bei der Sol-Gel Synthese von Aluminiumalkoxidfluorid, diploma thesis, Humboldt University of Berlin (2006). [17] A. Abraham, R. Prins, J. A. van Bokhoven, E. R. H. van Eck and A. P. M. Kentgens, Multinuclear solid-state high-resolution and 13C-{27Al} double-resonance magic-angle spinning NMR studies on aluminium alkoxides, J. Phys. Chem. B, 110, 6553–6560 (2006).

36

Functionalized Inorganic Fluorides

[18] V. J. J. Shiner, D. Whittaker and V. P. Fernandez, The structures of some aluminum alkoxides, J. Am. Chem. Soc., 85, 2318–2322 (1963). [19] D. C. Kleinschmidt, V. J. J. Shiner and D. Whittaker, Interconversion reactions of aluminum isopropoxide polymers, J. Org. Chem. 38, 3334–3337 (1973). [20] J. W. Akitt and R. H. Duncan, Multinuclear study of aluminium isopropoxide – Fourier Transform NMR of a quadrupolar nucleus, J. Magn. Reson. 15, 162–165 (1974). [21] O. Kriz, B. Casensky, A. Lycka, J. Fusek and S. Hermanek, 27Al NMR behaviour of aluminium alkoxides, J. Magn. Reson., 60, 375–381 (1984). [22] N. Ropson, P. Dubois, R. Jerome and P. Teyssie, Macromolecular Engineering of polylactones and polalactides. A 13C and 27Al NMR study of the effect of g- Butyrolactone on the structure of Aluminium isopropoxide in toluene, Macromol., 26, 6378–6385 (1993). [23] R. Ko¨nig, G. Scholz, N. H. Thong and E. Kemnitz, Local structural changes at the formation of fluoride sols and gels – a mechanistic investigation by multinuclear NMR spectroscopy, Chem.Mater. 19 (9), 2229–2237 (2007). [24] P. J. Chupas, D. R. Corbin, V. N. M. Rao, J. C. Hanson and C. P. Grey, A combined solid- state NMR and diffraction study of the structures and acidity of fluorinated alumina: implications for catalysis, J. Phys. Chem. B, 107, 8327–8336 (2003). [25] D. Dambournet, A. Demourgues, C. Martineau, J. Majimel, M. Feist, C. Legein, J. Y. Buzare´, F. Fayon and A. Tressaud, Nanostructured Al-based Fluoride-Oxide Materials with a Core-Shell Morphology, J. Phys. Chem. C, 112, 12374–12380 (2008). [26] R. Ko¨nig, G. Scholz and E. Kemnitz, Crystalline aluminium hydroxy fluorides – suitable reference compounds for 19F chemical shift trend analysis of related amorphous solids, J. Fluorine Chem., 129, 598–606 (2008). [27] R. Ko¨nig, G. Scholz, A. Pawlik, C. Ja¨ger, B. v. Rossum, H. Oschkinat and E. Kemnitz, Crystalline aluminium hydroxy fluorides: structural insights obtained by high-field solid state NMR and trend analyses, J. Phys. Chem. C, 112, 15708–15720 (2008). [28] P. J. Chupas, M. F. Ciraolo, J. C. Hanson and C. P. Grey, In situ X-ray diffraction and solid-state NMR study of the fluorination of g-Al2O3 with HCF2Cl, J. Am. Chem. Soc., 123, 1694–1702 (2001). [29] E. Kemnitz, Sol-gel-synthesis of metal fluorides – a new tool for fluoride based materials with exciting properties, 18th Winter Fluorine Conference, St. Pete Beach, Florida, USA, January 2007. [30] J. Koch, Untersuchungen zum Einfluß von Pyridin bei der Sol-Gel Synthese von HS-AlF3 und phosphonsa¨ure-modifiziertem Aluminiumlfuorid, diploma thesis, Humboldt University of Berlin, (2008). [31] A. Dimitrov, S. Wuttke, S. I. Troyanov and E. Kemnitz, Mg6F2(OMe)10(MeOH)14 – an alkoxide fluoride of an alkaline earth metal, Angew. Chem. Int. Ed., 47, 190–192 (2008). [32] S. Wuttke, A. Lehmann, G. Scholz, M. Feist, A. Dimitrov, S. I. Troyanov and E. Kemnitz, Investigation of the fluorolysis of magnesium methoxide, Dalton Trans., 4729–4734 (2009). [33] Z. A. Starikova, A. I. Yanovsky, E. P. Turevskaya and N. Ya. Turova, The structure of the crystal solvate of magnesium methoxide with methanol (OMe)23.5MeOH, Polyhedron 16, 967–974 (1997). [34] R. Ko¨nig, G. Scholz and E. Kemnitz, Local structural changes in aluminium isopropoxide fluoride xerogels and solids as consequence of progressive fluorination degree, J. Phys. Chem. C, 113, 6426–6438 (2009). [35] R. Ko¨nig, G. Scholz, A. Pawlik, C. Ja¨ger and E. Kemnitz, Identification of AlFx(OR)y Species in strongly disordered Aluminium Isopropoxide Fluoride Solids: A field-dependent MAS NMR study, J. Phys. Chem. C, 113, 15576–15585 (2009). [36] D. Dambournet, A. Demourgues, C. Martineau, S. Pechev, J. Lhoste, J. Majimel, A. Vimont, J.-C. Lavalley, C. Legein, J.-Y. Buzare, F. Fayon, A. Tressaud, Nanostructured aluminium hydroxyfluorides derived from -AlF3, Chem. Mater, 20, 1459–1469 (2008). [37] D. Dambournet, A. Demourgues, C. Martineau, E. Durand, J. Majimel, A. Vimont, H. Leclerc, J.-C. Lavalley, M. Daturi, C. Legein, J. Y. Buzare´, F. Fayon and A. Tressaud, Structural investigation and acidic properties of high surface area pyrochlore aluminium hydroxyfluoride, J. Mater. Chem.18, 2483–2492 (2008).

Sol-Gel Synthesis of Nano-Scaled Metal Fluorides – Mechanism and Properties

37

[38] R. Ko¨nig, G. Scholz and E. Kemnitz, New inserts and low temperature: two strategies to overcome the bottleneck in MAS NMR on wet gels, Solid State NMR, 32, 78–88 (2007). [39] D. R. Neuville, L. Cormier and D. Massiot, Al environment in tectosilicate and peraluminous glasses: A 27Al MQ-MAS NMR, Raman and XANES investigation, Geochim. et Cosmochim. Acta, 68, 5071–5079 (2004). [40] A. Pawlik, R. Ko¨nig, G. Scholz, E. Kemnitz, G. Brunklaus, M. Bertmer and C. Ja¨ger, Access to local structures of HS-AlF3 and its precursor determined by high resolution solid-state NMR, J. Phys. Chem. C, 113, 16674–16680 (2009). [41] A. Pawlik, R. Ko¨nig, G. Scholz, E. Kemnitz and C. Ja¨ger, HS-AlF3 and its precursor: the structure of amorphous systems investigated by 19F and 1H solid state NMR, in preparation, J. Phys. Chem. C, (2010). [42] U. Groß, St. Ru¨diger, E. Kemnitz, K.-W. Brzezinka, S. Mukhopadhyay, C. L. Bailey, A. Wander and N. M. Harrison, Vibrational analysis study of aluminium trifluoride phases, J. Phys. Chem. A, 111, 5813–5819 (2007). [43] S. Chaudhuri, P. Chupas, B. J. Morgan, P. A. Madden and C. Grey, An atomistic MD simulation and pair-distribution-function study of disorder and reactivity of -AlF3 nanoparticles, Phys. Chem. Chem. Phys., 8, 5045–5055 (2006). [44] R. Ko¨nig, Lokale Strukturen nanoskopischer Aluminiumalkoxidfluoride und chemisch verwandter kristalliner Verbindungen, PhD thesis, Humboldt University of Berlin, (2009). [45] J. K. Murthy, U. Groß, St. Ru¨diger, V. V. Rao, V. V. Kumar, A. Wander, C. L. Bailey, N. M. Harrison and E. Kemnitz, Aluminium chloride as a solid is not a strong Lewis acid, J. Phys. Chem. B, 110, 8314–8319 (2006). [46] K. O. Christe, D. A. Dixon, D. McLemore, W. W. Wilson, J. A. Sheehy and D. A. Boatz, On a quantitative scale for Lewis acidity and recent progress in polynitrogen chemistry, J. Fluorine Chem., 101, 151–153 (2000). [47] Th. Krahl. A. Vimont, G. Eltanany, M. Daturi and E. Kemnitz, Determination of the acidity of high surface AlF3 by IR spectroscopy of adsorbed CO probe molecules, J. Phys. Chem. C, 111, 18317–18325 (2007). [48] M. Ahrens, G. Scholz and E. Kemnitz, Synthesis and crystal structure of RbKLiAlF6 – the first Al-Elpasolite with three different alkali metals, Z. Anorg. Allg. Chem. 634, 2978–2981 (2008). [49] U. Groß, St. Ru¨diger and E. Kemnitz, Alkaline earth fluorides and their complexes: A sol-gel fluorination study, Solid State Sci. 9, 838–842 (2007). [50] K. Tanabe, T. Sumiyoshi, K. Shibata, T. Kiyoura and J. Kitagawa, A new hypothesis regarding the surface acidity of binary metal oxides, Bull. Chem. Soc. Japan, 47, 1064–1066 (1974). [51] E. Kemnitz, Y. Zhu and B. Adamczyk, Enhanced Lewis acidity by aliovalent cation doping in metal fluorides, J. Fluorine Chem., 114, 163–170 (2002). ¨ nveren and E. Kemnitz, Mixed metal fluorides as [52] J. Krishna Murthy, U. Groß, St. Ru¨diger, E. U doped Lewis acidic catalysts systems: a comparative study involving novel high surface area metal fluorides, J. Fluorine Chem., 125, 937–949 (2004). [53] J. Krishna Murthy, U. Groß, St. Ru¨diger and E. Kemnitz, FeF3/MgF2: novel Lewis acidic catalyst systems, Appl. Catal., A, 278, 133–138 (2004). ¨ nveren, W. Unger and E. Kemnitz, Synthesis and [54] J. Krishna Murthy, U. Groß, St. Ru¨diger, E.U characterization of chromium(III)-doped magnesium fluoride catalysts, Appl. Catal., A., 282, 85–91 (2005). [55] M. Ahrens, G. Scholz, M. Feist and E. Kemnitz, Application of an alkoxide sol-gel route for the preparation of complex fluorides of the MAlF4 (M ¼ K, Cs), M3AlF6 (M ¼ Li, Na, K), and Na5Al3F14 type, Solid State Sci., 8, 798–806 (2006). [56] M. Ahrens, K. Schuschke, S. Redmer and E. Kemnitz, Transparent ceramics from sol-gel derived elpasolites by cold pressing, Solid State Sci., 9, 833–837 (2007). [57] St. Wuttke, S. M. Coman, G. Scholz, H. Kirmse, A. Vimont, M. Daturi, S. L. M. Schroeder and E. Kemnitz, Novel Sol-Gel Synthesis of acidic MgF2-x(OH)x materials, Chem. Eur. J. 14, 11488–11499 (2008). [58] S. M. Coman, S. Wuttke, A. Vimont, M. Daturi and E. Kemnitz, Catalytic performance of nanoscopic - AlF3 based catalysts in the synthesis of (all-rac)--tocopherol, Adv. Synth. Catal., 350, 2517–2524 (2008).

38

Functionalized Inorganic Fluorides

[59] S. M. Coman, P. Patil, St. Wuttke and E. Kemnitz, Cyclisation of citronellal over heterogeneous inorganic fluoride – highly chemo- and diasterioselective catalysts for (–)-isopulegol, Chem. Commun. 460–462 (2009). [60] K. Scheurell and E. Kemnitz, Amorphous aluminium fluoride as new matrix for vanadium containing catalysts, J. Mater. Chem., 15, 4845–4853 (2005). [61] Pratap T. Patil, A. Dimitrov, J. Radnik and E. Kemnitz, Sol-gel synthesis of metal fluorides supported Pd catalysts for Suzuki coupling, J. Mater. Chem., 18, 1632–1635 (2008). [62] G. Li, L. Wang, H. Ni and C. U. Pittmann, Jr., Polyhedral Oligomeric Silsesquioxane (POSS) Polymers and Copolymers: A Review, J. Inorg. Organomet. Polym. 11, 123–154 (2002). [63] S. Ku¨hl, Organisch modifizierte Metallfluoride – anorganisch-organische Hybridsysteme durch Sol-Gel-Synthese, diploma thesis, Humboldt University of Berlin, (2007). [64] G. Eltanany, St. Rudiger and E. Kemnitz, Supported high surface AlF3: a very strong solid Lewis acid for catalytic applications, J. Mater. Chem., 18, 2268–2275 (2008). [65] T. Murata, H. Ishizawa, I. Motoyama and A. Tanaka, Praparation of high-performance optical coatings with fluoride nanoparticle films made from autoclaved sols, Applied Optics 45, 1465–1468 (2006). [66] H. Kru¨ger, E. Kemnitz, A. Hertwig and U. Beck, Transparent MgF2-films by sol-gel coating: synthesis and optical properties, Thin Solid Films 516, 4175–4177 (2008). [67] H. Kru¨ger, E. Kemnitz, A. Hertwig and U. Beck, Moderate temperature sol-gel deposition of magnesium fluoride films for optical applications: A study on homogeneity using spectroscopic ellipsometry, Phys. Stat. Sol. 205, 821–824 (2008).

2 Microwave-Assisted Route Towards Fluorinated Nanomaterials Damien Dambournet, Alain Demourgues and Alain Tressaud Institut de Chimie de la Matie`re Condense´e de Bordeaux (ICMCB-CNRS), Universite´ Bordeaux 1. 87 Avenue du Dr Albert Schweitzer, 33608 Pessac Cedex, France

2.1

Introduction

The development of synthesis routes for the preparation of new inorganic fluorides is of continuing interest. There are numerous routes for synthesizing metal fluorides that include solid-state reactions, aqueous and sol-gel synthesis, gas-phase reactions and decomposition reactions.1,2 Since the beginning of the millennium, the synthesis of nanomaterials has gained in importance due to the intrinsic size-dependent properties of the resulting solids. In the field of synthesis techniques, the use of microwaves as a heating source is a growing area due to its peculiar heating mode.3 In the traditional heating mode, heat is transferred from the vessel toward the mixture. Microwave heating, however, takes place directly in the core of the matrix, allowing shorter reaction times and the absence of temperature gradient in the medium. It also offers the possibility of stabilizing new nanosized materials in an energy-efficient way. A large range of materials has been successfully prepared including oxides,4 phosphates,5 sulphides,6 hybrid fluorides,7 metals,8 nanoporous materials9 . . . However, the application of such a technique for preparing nanosized metal fluorides is very limited so far.

Functionalized Inorganic Fluorides: Synthesis, Characterization & Properties of Nanostructured Solids Edited by Alain Tressaud  2010 John Wiley & Sons, Ltd

40

Functionalized Inorganic Fluorides

The first part of this chapter undertakes an introduction to microwave synthesis. After presenting the origin of this technique, some basic concepts of microwave heating are described and the advantages offered by this route are briefly discussed. Finally, a typical microwave oven dedicated to the synthesis is shown. Some examples of nanosized metal fluorides prepared by microwave irradiation are presented in the subsequent section. Particular attention is given to the preparation of nanosized aluminium-based fluoride materials. The impact of some synthesis parameters is discussed and a detailed characterization of the Al-based compounds is proposed to highlight the potential of this preparation route.

2.2 2.2.1

Introduction to Microwave Synthesis A Brief History

The potential use of microwave irradiation as a heating source was unexpectedly discovered by Percy L. Spencer. In 1945, while working on the development of radar through the use of a magnetron, he discovered that a chocolate bar in his pocket had melted. Thereafter, he confirmed his discovery and proposed through a patent the use of the microwave irradiation to heat some foodstuffs.10 In the 1970s, the development of this technology led to the largescale use of domestic microwave ovens. In 1986, two papers11 reported that several organic reactions could be accelerated using microwave irradiation. Since then, the use of microwave as a heating source in chemistry has continued to grow regularly. Nevertheless, at an early stage of the method, the use of the domestic microwave oven without any temperature/ pressure controls has led to unsafe and irreproducible works.12 Fortunately, the growth of this technology has enabled, more recently, the development of microwave oven dedicated to the chemical synthesis, leading to safer works and reproducible results.

2.2.2

Mechanisms to Generate Heat

Microwaves are electromagnetic radiations lying between radio wave frequencies and infrared frequencies (between 0.3 and 300 GHz). These are produced by a magnetron, which consists of a thermionic diode having an anode and a directly heated cathode.13 Microwaves contain an electric and a magnetic field component. It is the interaction between the electric field component and the matter that generates the heat through two mechanisms.13 (i) Dipolar polarization. A molecule that possesses a dipolar moment, such as water, is sensitive to an electric field. Under the influence of the latter, the dipole will attempt to align itself by rotation. The applied field provides the energy for this rotation, which depends on the frequency. The rapid direction change of the microwave components cannot be followed by the dipole. Therefore, a phase difference between the orientation of the field and that of the dipole is generated. This phase difference creates energy, which is lost from the dipole by molecular friction and collisions, giving rise to heat (dielectric heating).

Microwave-Assisted Route Towards Fluorinated Nanomaterials

41

The frequencies that allow microwave dielectric heating to take place are 918 MHz and 2.45 GHz, the latter being the most used. This mechanism depends on the ability of the dipole to reorientate under the applied electric field. The capability of a substance to convert the electromagnetic energy into heat is given by the dielectric loss: tan  that is equal to the ratio of the dielectric loss "00 to the relative permittivity "0 . The relative permittivity represents the measure of the ability of a molecule to be polarized, while the dielectric loss is the ability of a medium to convert dielectric energy into heat. (ii) Ionic conduction. In a solution, the ions will move under the influence of an electric field component. As a consequence, the collision rate increases, converting the kinetic energy into heat. This phenomenon explains why the temperature of two samples containing distilled water and tap water will be higher in the case of tap water after similar microwave irradiation conditions.14

2.2.3

Advantages of Microwave Synthesis

Microwave synthesis offers a large range of advantages over conventional synthesis. The consequences of the direct heating are the following: reduction of energy consumption, better yield and gain in the reaction duration. From a practical viewpoint, the shorter reaction time enables a faster optimization of the synthesis parameters. The ‘microwave chemistry’ can be applied to practically all inorganic families that do exist (see Introduction). Based on the literature, some interesting facts have been reported and can be summarized as follows: stabilization of metastable and novel phases, increase of the phase purity and phase selectivity, short crystallization times and narrow particle-size distribution. The explanations for these advantages are still poorly understood but several reasons9 are usually pointed out: very rapid heating rate, uniform heating, unusual interactions between species in the reaction mixture, occurrence of hot spots and enhancement of the dissolution of the reacting species. The occurrence of specific microwave effects is still under debate to explain the whole advantages displayed by this technique. For instance, an hypothesis based on enhanced diffusion mechanisms has been recently proposed.15 Some breakthroughs should be expected to occur with the development of in situ techniques.9

2.2.4

Examples of Microwave Experiments

As mentioned above, the development of microwave-oven engineering has allowed a drastic enhancement of safety and reproducibility. Several ovens dedicated to chemistry are now available. For further details, the reader can referred to the paper from Kremsner et al.16 In the second part of this chapter, different compounds have been synthesized using an advanced microwave oven from the CEM Corporation (Figure 2.1). The oven is a MARS-5 Microwave Digestion System operating at a frequency of 2.45 GHz and a maximum power of 1200 W. The reactor used (XP-1500 plus model) can operate up to 220 C and 55 bars as internal pressure. To prevent explosion, any deviation in the predetermined parameters causes the system to stop. Secondly, in case of overpressure, the breaking of a disk occurs and the released gaseous species are evacuated by the

42

Functionalized Inorganic Fluorides

ventilation system. The temperature is regulated by percentage increments of the microwave power and controlled by an optical fiber. Internal pressure is measured by a pressure sensor. Several parameters can be tuned: temperature, pressure, time, heating rate and microwave power. In a typical synthesis procedure, the precursor solution is placed in a Teflon container that is fixed onto a rotating device (Figure 2.1, left). After cooling down to room temperature, the reactor is open and set in a second device that allows the solvent evaporation (right). This final step is performed under primary vacuum and argon flow.

Figure 2.1 Images of a microwave oven dedicated to chemical reactions

2.3 2.3.1

Preparation of Nanosized Metal Fluorides Aluminium-based Fluoride Materials

Numerous studies have been devoted to the investigation of aluminium fluorides. Owing to the affinity17 of fluoride ions toward Al3þ and the strong electronegativity of F ions, the Al-F system gives rise to numerous crystallographic forms.18–23 In a general way, aluminium fluorides are obtained by thermal decomposition of a precursor, which determines the final structural arrangement.24 Scheme 2.1 summarizes the synthesis of the various forms of AlF3 and their transformation toward the thermodynamically stable -AlF3 variety (ReO3-derived type).25 Aluminium fluorides were reported to be particularly suitable catalysts in the synthesis of CFCs substitutes. Recall that, owing to the strong stability of the F–C bond as compared to those of fluorine with other elements (N, Cl, Br, H . . .)24 organic fluorine chemistry has been strongly developed,26, 27 leading, for instance, to the worldwide use of chlorofluorocarbons (CFCs) in industry (refrigerants, aerosol, solvents . . .). The role of these compounds in the ozone-depleting phenomenon28 led to a drastic rise in the research into heterogeneous catalysts that would be able to convert the CFCs into more environmentally acceptable compounds – hydrogenfluorocarbons

Microwave-Assisted Route Towards Fluorinated Nanomaterials

43

(HFCs).29 In the Al-F system, the hexagonal tungsten bronze type structure, labelled HTB or -AlF3, is the most known compound because of its high catalytic properties toward halogen exchange reaction.30 Beside efforts to understand their catalytic activity, synthetic routes were developed leading to new metastable forms of AlF3. Herron et al.23, 31 developed a nonaqueous route to prepare fluoroaluminate salts with general formula MAlF4, where M is an organic cation (M ¼ pyridineHþ, N(CH3)4þ, NH4þ). The thermal decomposition of these salts leads to new crystalline AlF3 through the release of MF gaseous species (Scheme 2.1). As far as the catalytic properties are concerned, the reactivity of the aluminium fluorides is also governed by the specific surface area displayed by the solid. Because of the strong electronegativity and reactivity of fluoride ions, conventional syntheses of AlF3 result in low surface area ( 3, besides the HTB phase, the thermodynamically stable phase19 -AlF3 appears and became predominant with increasing HF contents. Nevertheless, in these conditions the -AlF3 phase cannot be prepared as a single phase, the HTB phase being always detected even for large HF contents. 2.3.1.1.2 Effect of the Nature of the Aluminium Precursor The nature of the aluminium precursor has been proved to affect both the nature of the stabilized phase and its morphology.

Microwave-Assisted Route Towards Fluorinated Nanomaterials

45

R = 3.5 HTB + α-AlF3

R=3 HTB

R = 2.8 Pyrochlore + HTB R=2 Pyrochlore 10

20

30

40 50 2θ (°)

60

70

80

Figure 2.2 Effect of the R ¼ [HF]/[Al] molar ratio on the X-ray diffraction powder pattern. For each X-ray diagram, the R molar ratio is noticed as well as the stabilized phase. Syntheses were conducted at T ¼ 160 C, t ¼ 2 h using Al(NO3)3.9 H2O and water/isopropanol as solvents

Influence on the Nature of the Stabilized Form For R ¼ 3, the use of an Al nitrate precursor favours the formation of the HTB phase while the use of chloride leads to a phase mixture containing -AlF3 as the major phase.41 The stabilization of the HTB phase is in fact related to a microwave-induced side reaction. There are several reports on the ability of microwave irradiation to induce side reactions such as condensation of alcohol molecules,44 decomposition of organic molecules45 and ionic liquids.46 When using nitrate as precursor, an exothermic effect is observed during the synthesis leading a drastic increase of both pressure and temperature. Such a reaction implies a redox process identified by chemical analysis, as the reduction of nitrate into ammonium ions coupled with the oxidation of isopropanol into ketones. The occurrence of NH4þ in the solid can be identified by FTIR spectroscopy. An increase in the synthesis temperature up to 170 C also confirmed the reduction of nitrate through the XRD identification of -NH4AlF4 as a side product. By analogy with the conventional route to HTB compounds, that is the decomposition of an hydrate or ammonium fluoride salts,39 it can be suggested that the ammonium ions act as a template for the stabilization of the phase. Another example of the effect of the cationic precursor on the stabilized form is displayed in Figure 2.3. For high HF contents (R ¼ 3.5), the use of nitrate as aluminium precursor leads to the stabilization of a phase mixture. In similar conditions, R ¼ 3.5, the use of aluminium chloride enables to get a single phase which XRD diagram is related to the ReO3 type structure (Figure 2.3). Influence on the Morphology The synthesis of the pyrochlore hydroxyfluoride using nitrate as precursor leads to the preparation of very well crystallized compounds with the drawback of a very low specific surface area (6 m2.g1). By replacing nitrate by

46

Functionalized Inorganic Fluorides AlCl3.9H2O ReO3 Phase type

Al(NO3)3.9H2O HTB +α-AlF3

10

20

30

40

50

60

70

80

2θ (°)

Figure 2.3 Influence of the aluminium precursor on the X-ray diffraction powder pattern for R ¼ 3.5. For each X-ray diagram, the precursor and the final form are noted

isopropoxide, a broadening of the X-ray peaks is observed (Figure 2.4) suggesting a lowering of the crystallite size. This trend is confirmed by the increase of the specific surface area, from 6 to 77 m2.g1. The crystallization growth of pyrochlore crystallites is thus strongly affected by the nature the aluminium precursor: the weakening of the crystallization growth could be due to the steric hindrance of the isopropoxide ligand. 2.3.1.1.3 Effect of the Solvents The nature of the solvents is very important, affecting both the morphology of the solid and its chemical composition. At first, two examples highlight the impact of the solvent when

Al(OiPr)3 S = 77 m2.g−1

Al(NO3)3.9H2O S = 6 m2.g−1

10

20

30

40

50

60

70

80

2θ (°)

Figure 2.4 Dependence of the X-ray diffraction powder pattern of the pyrochlore hydroxyfluoride on the nature of the aluminium precursor. For each X-ray diagram, the precursor and the specific surface area is noted

Microwave-Assisted Route Towards Fluorinated Nanomaterials

47

targeting high surface area materials. In the case of the pyrochlore hydroxyfluoride, the addition of a small amount of ether drastically improves the surface area of the solid, from 77 to 140 m2.g1.42 Following the work of Kemnitz et al.,33 ether molecule in the presence of HF has been suggested to form an oxonium ion giving rise to ion pairs with F. Such a complex should lead to a decrease of the fluorine reactivity known to be a structuring agent and therefore enable the preparation of nanosized crystallites. During the synthesis of the HTB-type compound, the surface area could be monitored by tuning the volume ratio V ¼ Vwater/Visopropanol of the used solvents. For V ¼ 1, a surface area of 82 m2.g1 is obtained, whereas for V ¼ 6, the value drops to 3 m2.g1. Interestingly, both compounds not only differ in their morphology but also by their chemical composition. Using elemental analysis and high field 27Al NMR spectroscopy, the chemical composition of the low and high surface area compounds has been accurately determined as AlF2.4(OH)0.6 and AlF2.6(OH)0.4, respectively. Water thus favours the stabilization of OH groups inside the structure while isopropanol enables the stabilization of F ions. This trend is confirmed when using V>1, the powder X-ray diffraction pattern of the corresponding solid displaying some traces of -AlF3, that is a more fluorinated compound. 2.3.1.1.4 Coupling Sol-gel Alkoxy-fluoride Route and Microwave Irradiation A more sophisticated route has been proposed by combining nonaqueous sol-gel fluoride synthesis and microwave solvothermal process.47 The nonaqueous sol-gel process developed by Kemnitz et al.,33 enables the preparation of X-ray amorphous metal fluorides through two steps. First, an alcoholic solution of a metal alkoxide is partially fluorinated by anhydrous hydrogen fluoride previously dissolved in an organic solvent through the reaction: M(OR)3 þ xHF ! MFx(OR)3-x þ xROH.48 After drying, the resulting alkoxy-fluoride is fluorinated using gaseous CFCs or anhydrous HF, leading to X-ray amorphous metal fluorides with very high surface area. In a second step, the alkoxy-fluoride sol-gel is subjected to a microwave solvothermal process at various temperatures ranging from 90 to 200 C for 1 hour treatment. Finally, the mixture is dried under microwave irradiation but this step has no impact on the morphology of the final compound. The characteristics of the dry gel obtained by microwave synthesis are gathered in Table 2.3. While low-temperature microwave treatment (10 ppm ascribed to the partial decomposition of the pyrochlore has not been mentioned Line

iso (–0.5)

 Q (–5)

1 2 3 4

–12.5 –8.7 –4.2 1.8

345 540 625 505

1 2 3

–15.5 –11.7 –9.5

280 610 990

Relative Intensity (–0.5)

Assignment of the 27Al NMR lines

Pyrochlore AlF1.7(OH)1.3 47.3 AlF6, AlF5(OH) and AlF4(OH)2 33.4 AlF3(OH)3 15.5 AlF2(OH)4 3.9 AlF(OH)5 -AlF2.6(OH)0.4 82 16 2

AlF6 and AlF5(OH) AlF4(OH)2 AlF3(OH)3

Microwave-Assisted Route Towards Fluorinated Nanomaterials

57

Pyrochlore AlF1.7(OH)1.3 2

3

1

β-AlF2.6(OH)0.4

84

82

80

78

76

74

72

70

68

BE (eV)

Figure 2.11 Al 2 p XPS spectra of the pyrochlore and HTB-type compounds Table 2.8 Al 2 p XPS data of the HTB and the pyrochlore phases obtained by spectrum reconstruction using FWHM of 2.1 eV. An alignment between the components of both compounds has been undertaken with 0.1 eV as a standard deviation to overcome the reference issue Line Label

1 2 3

Al 2 p position eV (–0.1)

74.4 76 77.4

Relative intensity (%) Pyrochlore

HTB

16 73.5 10.5

22 54 24

A New Form of Aluminium Fluoride Hydrate It has been noted (cf. 2.3.1.1.2) that using chloride as aluminium precursor, a single phase whose XRD is closely related to the hightemperature phase of -AlF3 could be obtained (Figure 2.12). It can be added that using a higher synthesis temperature, that is 180 C instead of 140 C, an additional X-ray peak is detected, suggesting the presence of -AlF3 (113) line (Inset Figure 2.12). This point highlights the need for a careful preparation to achieve phase purity. Using the International Centre for Diffraction Data (ICDD), the powder pattern could be indexed in a first step as a cubic phase ˚ ) with the chemical formula AlF3.H2O.60 Nevertheless, the structural (Pm-3 m, a ¼ 3.600 A description appeared unrealistic because of a poor reliability factor (RBragg ¼ 16.4 %). The structure of this new compound has recently been investigated.43 At first, FTIR

58

Functionalized Inorganic Fluorides

spectroscopy and thermogravimetric analysis confirmed that the prepared compound consists indeed of a hydrated aluminium fluoride, but the attempts to place water molecule in the conventional cubic cell led to a poor agreement factor (RBragg  18 %). Water molecules were finally considered as part of the first coordination sphere of Al3þ ions. Nevertheless, because water molecule is neutral, [AlF6x(H2O)x]x3 octahedra display a lack of electron density, which leads to consider the occurrence of cationic vacancies. The latter has been clearly demonstrated by the convergence of the reliability factor down to RBragg ¼ 4.3 % when refining the occupancy rate of the Al3þ atoms and introducing water in the anionic fluorinated site (3a). The chemical formula Al0.82&0.18F2.46(H2O)0.54 could be proposed based on elemental analysis, TGA and Rietveld refinement. The crystallographic data of Al0.82&0.18F2.46(H2O)0.54 are gathered in Table 2.9 and compared to those of -AlF3. Interestingly, the interatomic distances for both hydrate and -AlF3 are very close. While the occurrence of the water molecule should lead to an increase of the interatomic distances as found for other aluminium fluoride hydrate,61 cationic vacancies may act as a repulsive entity decreasing the Al-X distances and maintaining the cubic symmetry. Table 2.9 Crystallographic data of Al0.82&0.18F2.46(H2O)0.54, a-AlF3 and the high temperature cubic form Phase

Crystal symmetry Space Group Z

Unit cell para- Interatomic ˚) ˚ )Al-X meters (A distances(A

Al0.82&0.18F2.46(H2O)0.54 Cubic Pm-3 m 1 Rhombohedral R-3 c -AlF3 12 High temperature cubic Cubic Pm-3 m 1 form

3.6067(1) 4.9305(6) 12.4462(7) 3.58

1.8034(1) 1.797(3)

Angles () 180 157.07(7) 180

1.791

T=180°C

41

20

30

40

50

43

60

45

70

80

Figure 2.12 X-ray diffraction pattern of the aluminium fluoride hydrate obtained after microwave irradiation at T ¼ 140 C. Inset: effect of the temperature on the XRD

Microwave-Assisted Route Towards Fluorinated Nanomaterials

59

In both hydroxyfluoride and hydrate compounds, the anionic sites are statically occupied either by F, OH ions or H2O molecules, thus generating disorder. Refinement of the Debye-Weller factors enables to get information about the disordering state in the anionic sites. A comparison between the three phases is displayed in Table 2.10. The hydroxyfluoride phases clearly show the impact of the OH/F substitution on the ordering state of the anionic sites, an increase of the substitution leading to higher Debye-Weller factors. While the substitution rate (F/H2O) in the anionic site is rather low, the hydrate phase Al0.82&0.18F2.46(H2O)0.54 exhibits the highest disorder. FTIR and 1H MAS NMR spectroscopy indicate that water molecule is in average located on the 3 d Wyckoff position inducing the large thermal displacement observed. The local structure has been characterized using 19F and 27Al MAS NMR (Figure 2.13). By analogy with the isotropic chemical shift of the hydroxyl-fluorinated species found in the HTB and pyrochlore phases, the AlF6 and AlF6x(H2O)x species (x ¼ 1, 2, 3) have been identified and quantified (Table 2.11). Interestingly, two contributions, reconstructed using four signals (Table 2.11), are displayed by the 19F MAS NMR spectrum (Figure 2.13). The major one ( 88 %) lies in the range of fluoride ions as found in -AlF3 and in HTB and pyrochlore hydroxyfluorides and is therefore ascribed to bridging fluorine atoms in AlF6x(H2O)x octahedra. The minor contribution (12 %) exhibits chemical shift values as observed in AlF3.9 H2O (–149.5 ppm) and -AlF3.3 H2O (–147.9 ppm), which are both built from isolated octahedra.61 The resulting fluoride atoms are thus assigned to nonbridging fluorine atoms in AlF6x(H2O)x octahedra and therefore localized next to an aluminium vacancy. The low contribution of nonbridging fluorine atoms gives the evidence that the aluminium vacancy is mostly surrounded by water molecules. FTIR spectroscopy gives information about the conformation of water molecules inside the structure. A 2:1 complex has been detected by FTIR, which can be noted as XH-O-HX where X is a proton acceptor (X ¼ F–, H2O). By comparison with IR bands of water in an inert solvent,62 the vibration bands of OH groups of water molecules are shifted due to hydrogen bonding towards lower wavenumber that is around 3185 and 3335 cm1 for  sym(OH) and  asym(OH), respectively. A representation of the structure is presented in Figure 2.14. 2.3.1.3

Key Points and Concluding Remarks

The above mentioned studies on aluminium-based fluorides compounds prepared by microwave-assisted synthesis have enabled the improvement of the knowledge on these materials as well as the development of different compounds having unusual properties in Table 2.10 Comparison of the Debye-Weller factors obtained from Rietveld refinement for the Al-based fluorides prepared by microwave irradiation ˚ 2) Biso (A

Phase

Pyrochlore AlF1.7(OH)1.3 HTB AlF2.4(OH)0.6 HTB AlF2.6(OH)0.4 Al0.82&0.18F2.46(H2O)0.54

Al3þ

X (X ¼ F, OH, H2O)

0.78(2) 0.62(3) 0.79(3) 0.6(3)

1.12(2) 1.25(4) 0.83(3) 2.1(4)

Functionalized Inorganic Fluorides

60

Table 2.11 Al0.82&0.18F2.46H2O0.54 line label, isotropic chemical shift diso (ppm), quadrupolar product  Q (kHz), relative line intensity (%) deduced from the reconstruction of the 27Al MAS NMR spectrum and the corresponding assignments. Isotropic chemical shift diso and relative intensity deduced from the reconstruction of the 19F MAS NMR spectrum 27

Line 1 2 3 4

Al NMR

iso (–0.5)

 Q (–5)

Relative Intensity (–0.5)

Assignment

17.9 14.4 11.7 –9.4

118 252 360 559

59.1 28.7 7.8 4.4

AlF6 AlF5(H2O) AlF4(H2O)2 AlF3(H2O)3

Relative Intensity (–0.5)

Assignment

13.4 74.3 4.4 7.9

Bridging F

19

F NMR

iso(–0.5) 173.0 169.4 155.2 152.5

Non-bridging F

1 2

3 4

−140

−160 19F δ iso

−180

−200

10

0

−10 27 Al δ iso

(ppm)

−20

−30

−40

(ppm)

2θ (°)

Figure 2.13 19F and 27Al MAS NMR spectra of Al0.82&0.18F2.46(H2O)0.54. (Reprinted with permission from [43] Copyright (2008) American Chemical Society.)

terms of morphology, composition and structure. The main highlights brought by the use of the microwave synthesis are summarized below: • A strong dependence of the [HF] content on the structural arrangement has been found to occur. Such a phenomenon has been rationalized by the structural impact of the hydroxyl groups. The coupling of sol-gel and microwave synthesis has also shown the effect of the temperature on the kinetics of fluorination. • Microwave-assisted synthesis has enabled the direct stabilization of metastable phases such as -AlF3. Additionally, it has enabled the preparation of a new phase with unusual structural features that is a derived form of -AlF3 containing cationic vacancies owing to the occurrence of water molecules in the vicinity of the cation.

Microwave-Assisted Route Towards Fluorinated Nanomaterials

61

• All the prepared materials exhibit very high surface area with nanosized crystallites as opposed to compounds prepared by conventional methods. • All the syntheses presented in this part have been obtained in one or two hours, which is unconventionally faster than comparable methods such as solvo-hydrothermal treatment. This fast and energy-efficient synthesis enabled a quicker investigation of the relevant parameters. Al B F F/OH2

A

2:1 complex

Al3+ vac

Isolated H 2O

2.55 Å

Figure 2.14 Representation of Al0.82&0.18F2.46(H2O)0.54: (A) view of the cubic cell. (B) Representation of the structure with an isolated water molecule. (Reprinted with permission from [43] Copyright (2008) American Chemical Society.)

2.3.2 2.3.2.1

Microwave-assisted Synthesis of Transition Metal Oxy-Hydroxy-Fluorides Titanium Oxy-hydroxy-fluorides

2.3.2.1.1 Influence of the [HF]/[Ti] Molar Ratio Like Al-based fluorides, the [HF] content expressed as the R ¼ [HF]/[Ti] molar ratio, is a decisive parameter for the final crystallized forms. Using oxychloride (TiOCl2.xHCl) as titanium precursor, the variation of the [HF] content leads to the stabilization of three phases (Figure 2.15): (i) for R ¼ 0.2, the anatase-type structure, (ii) for R ¼ 1.7, a new phase adopting a derived-HTB type structure and finally, (iii) for R ¼ 3 a superstructure of the ReO3 form. The two latter phases are obtained at T ¼ 90 C. On the other hand, the fluorine-doped anatase is obtained at higher temperature, that is T ¼ 150 C. The composition, structural features and UV-absorption properties of Ti oxy-hydroxy fluorides are detailed in chapter 8 of this book by A. Demourgues, L. Sronek and N. Penin.69–71 2.3.2.1.2 Impact on the Chemical Composition Here again, the use of microwave-assisted route enables the synthesis of a new series of Tibased fluoride nanomaterials at low duration time and low temperature. Beside, the report of a phase derived from the HTB-type structure, the synthesis conditions favour the competition between OH and F ions, leading to new compositions besides TiOF2, the only titanium oxyfluoride reported so far by various authors through various synthesis

62

Functionalized Inorganic Fluorides

R = 3/ReO3 form

R = 1.7/HTB form-derived

R = 0.2/F-anatase

10

20

30

40

50

60

70

80

2θ (°)

Figure 2.15 Dependence of the X-ray diffraction powder pattern on the R ¼ [HF]/[Ti] molar ratio. For each X-ray diagram, the R molar ratio is noticed as well as the nature of the stabilized phase. Syntheses were conducted at T ¼ 90 C, t ¼ 30 min (TiOCl2.xHCl in water), excepted for R ¼ 0.2: T ¼ 150 C t ¼ 30 min

approaches.63,64 On the contrary, the microwave process leads, for instance, to the stabilization of an hydroxyfluoride containing some cationic vacancies: Ti0.75F1.5(OH)1.5,65 which adopts a superstructure of the ReO3 form. Moreover, the HTB-derived form corresponds to the following formula Ti0.75O0.25(OH)1.3F1.2 with cationic vacancies and anionic nonstoichiometry, whereas the anatase form Ti0.9F0.2O1.6(OH)0.2 contains also three types of anions with cationic vacancies. 2.3.2.1.3 From Aqueous to Organic Medium Effect on the stabilized phase. The impact of the medium on the stabilization of Ti-based compounds has been investigated by tuning the solvent and cationic precursor while keeping the molar ratio R equal to 1.7. Figure 2.16 shows the X-ray diffraction powder patterns obtained using either aqueous (TiOCl2.xHCl in water) or organic media (Ti(OCH(CH3)2)4 in isopropanol). While the aqueous medium leads to a well-crystallized phase identified as a new oxy-hydroxy-fluoride, the use of isopropoxyde in isopropanol enables the preparation of fluorinated anatase. According to the width of the X-ray peaks, the solid as prepared is built of nanoparticles, which is confirmed by the HRTEM images showing crystallite dimensions of about 5 nm (Figure 2.17). Fast Fourier Transforms (FFT) of HRTEM image matched with a [111] zone axis of the I41/amd tetragonal anatase structure. Such a result is rather surprising keeping in mind the strong influence of the [HF] content on the stabilized form. The kinetics of fluorination is here clearly dependent of the nature of the ligand. 2.3.2.1.4 Tin Titanium Oxy-hydroxyfluoride Xie et al.,66 have synthesized a new tin-titanium oxy-hydroxy-fluoride Sn1.24Ti1.94O3.66(OH)1.50F1.42 exhibiting a pyrochlore-type structure. The synthesis

Microwave-Assisted Route Towards Fluorinated Nanomaterials

TiOCl2.xHCl H2O

10

20

30

50

40 2θ

60

63

Ti(OCH(CH3)2)4 Isopropanol

70

80

10

20

30

40

50

60

70

80

2θ

Figure 2.16 Dependence of the X-ray diffraction powder pattern oxy-hydroxy-fluoride on the medium used: aqueous (left), organic (right)

of

Ti-based

Figure 2.17 HRTEM image of F-anatase nanoparticules prepared using organic medium. Inset is the FFT of particle ˚ showing a [111] zone axis of the I41/amd tetragonal structure

conditions were optimized by tuning several parameters, i.e. the pH of the solution. The final solid is build of nanosized crystallites ranging from 20 to 30 nm. Interestingly, a rutile phase Sn0.39Ti0.61O2 with particle size around 100 nm is obtained by the decomposition of this oxy-hydroxy-fluoride at 800 C, which is far lower than the conventional route used generally for preparing this compound, i.e. 1450 C.

64

Functionalized Inorganic Fluorides

2.3.2.2

The Use of Ionic Liquids for the Synthesis of Metal Fluorides

In this case, metal fluorides are obtained by decomposition of ionic liquid. The latter is the source of F ions. The experiments have been performed in a domestic oven (the one that you are using to cook food).46 As a general method, a mixture containing the ionic liquid BMIBF4 (1-Butyl-3-methylimidazolium tetrafluoroborate) and a nitrate metal salt is subject to microwave irradiation in a domestic oven for a very short time, ranging from 5 to 10 min. In this way, the synthesis of various metal fluorides presenting nanosized and particular morphology were reported: FeF2 (nanorods), CoF2 (aggregated needles), ZnF2 (anisotropic morphology), LaF3 (oval) and YF3 (needle-shaped). The release of fluorine ions from the solvent was claimed to arise from an hydrolysis reaction such as BF4  þ H2 O ! BF3 :H2 O þ F . 2.3.2.3

Metal Organic Frameworks

In the search of porous inorganic-organics hydrides solids (see other contributions in this volume) which display several potential applications,67 the development of new synthesis routes is of crucial importance because the conventional methods generally consist of hydro- or solvo-thermal processes that require up to several days. The cubic chromium terephthalate Cr3F(H2O)2O[C6H4(CO2)2]3 (MIL-101), for instance, has been successfully synthesized using a microwave solvothermal process.68 Owing to the faster dissolution of the precursors and/or the condensation process, the MIL-101 can be obtained even after 1 min of irradiation at 210 C. Finally, a well crystallized material is obtained for 40 min of irradiation at 210 C while the conventional method is performed at 220 C for 10 h. The very rapid synthesis of MIL-101 leads to nanosized crystals that result in an improvement of the sorption properties of benzene.

2.4

Concluding Remarks

Although microwave ‘inorganic’ chemistry has been developed only recently, such a technique has already showed extraordinary potential. Its application to the preparation of metal fluorides clearly offers numerous advantages: the direct stabilization of metastable phase, the isolation of new phases, the preparation of homogenous and nanosized particles leading to high surface area materials, the easy control of the kinetic of fluorination. There are many applications of the prepared materials, including catalysis, optics and energy storage. Moreover, the unconventional heating mode enables fast reaction to occur, which renders this route one of the most energy efficient currently available. It is clear that interest in such a technique will continue to growth in the future with a need for a deeper understanding of the microwave chemistry.

Acknowledgements The EU is gratefully acknowledged for financial support through the Sixth Framework Programme (FUNFLUOS, Contract No. NMP3-CT-2004–5005575). The

Microwave-Assisted Route Towards Fluorinated Nanomaterials

65

fluorination experiments have been carried out with the help of Etienne Durand. Most NMR results have been obtained at Universite´ du Maine by Christophe Legein and Jean-Yves Buzare´.

References [1] J. Grannec and L. Lozano,in Inorganic Solid Fluorides, P. Hagenmuller (Ed.), Academic Press, Inc., Orlando, FL, Chapter 2, 1985. [2] K. Lutar, H. Borrmann and M. Leblanc, in Advanced Inorganic Fluorides: Synthesis, Characterization and Applications, T. NaKajima, B. Zemva and A. Tressaud (Eds), Elsevier Science, Chapter 2, 2000. [3] D. Adam, Microwave chemistry: Out of the kitchen. Nature, 421, 571–572 (2003). [4] I. Bilecka, I. Djerdj and M. Niederberger, One-minute synthesis of crystalline binary and ternary metal oxide nanoparticles, Chem. Com., 7, 886–888 (2008). [5] A. Murugan, T. Muraliganth, and A. Manthiram, Comparison of microwave assisted solvothermal and hydrothermal syntheses of LiFePO4/C nanocomposite cathodes for lithium ion batteries, J. Phys. Chem. C, 112, 14665–14671 (2008). [6] T. Rath, B. Kunert, R. Resel, G. Fritz-Popovski, R. Saf and G. Trimmel, Investigation of primary crystallite sizes in nanocrystalline ZnS powders: Comparison of microwave assisted with conventional synthesis routes, Inorg. Chem., 47, 3014–3022 (2008). [7] S. P. Thanh, F. Gaslain, M. Leblanc and V. Maisonneuve, Rapid synthesis of hybrid fluorides by microwave heating, J. Fluor. Chem. 101, 161–163 (2000). [8] M. Tsuji, M. Hashimoto, Y. Nishizawa, M. Kubokawa and T. Tsuji, Microwave-assisted synthesis of metallic nanostructures in solution, Chem. Eur. J., 11, 440–452 (2005). [9] G. A. Tompsett, W. C. Conner and K. S. Yngvesson, Microwave synthesis of nanoporous materials, Chem. Phys. Chem., 7, 296–319 (2006). [10] Percy L. Spencer, Method of Treating Foodstuffs, U.S. Patent 2,495,429, United States Patent Office (1950). [11] R. Gedye, F, Smith, F. Westaway, H. Ali, L. Baldisera, L. Laberge and J. Rousell, The use of microwave ovens for rapid organic synthesis, Tetrahedron Lett., 27, 279–282 (1986). R. J. Giguere, T. L. Bray, S. M. Duncan and G. Majetich, Application of commercial microwave ovens to organic synthesis, Tetrahedron Lett., 27, 4945–4948 (1986). [12] M. Nu¨chter, B. Ondruschka, W. Bonrath and A. Gum, Microwave assisted synthesis – a critical technology overview, Green Chem., 6, 128–141 (2004). [13] D. M. P. Mingos and D. R. Baghurst, Applications of microwave dielectric heating effects to synthetic problems in chemistry. Chem. Soc. Rev., 20, 1–47 (1991). [14] P. Lidstrom, J. Tierney, B. Wathey and J. Westman, Microwave assisted organic synthesis – a review, Tetrahedron, 57, 9225–9283 (2001). [15] C. Antonio and R. T. Deam, Can ‘microwave effects’ be explained by enhanced diffusion?, Phys. Chem. Chem. Phys., 9, 2976–2982 (2007). [16] J.M. Kremsner, A. Stadler, and C. O. Kappe, The scale-up of microwave-assisted organic synthesis, Top. Curr. Chem., 266, 233–278 (2006). [17] K.O. Christe, D. A. Dixon, D. McLemore, W. W. Wilson, J. A. Sheehy and J. A. Boatz, On a quantitative scale for Lewis acidity and recent progress in polynitrogen chemistry, J. Fluor. Chem., 101, 151–153 (2000). [18] J. Ravez, A. Mogus-Milankovic, J. P. Chaminade and P. Hagenmuller, Ferroelactic properties of AlF3, Mat. Res. Bull., 19, 1311 (1984). [19] P. Daniel, A. Bulou, M. Rousseau, J. Nouet, J. L. Fourquet, M. Leblanc and R. J. Burriel, A study of the structural phase-transitions in AlF3 - X-ray powder diffraction, DSC and Raman scattering investigations of the lattice-dynamics and phonon spectrum, Phys.: Condens. Matter, 2, 5663 (1990).

66

Functionalized Inorganic Fluorides

[20] J. L. Fourquet, M. Rivie`re and A. Le Bail, Crystal structure and protonic conductivity of pyrochlore phase Al2[(OH)1xFx]6.H2O and Al2[(OH)1xFx]6, Eur. J. Solid State Inorg. Chem., 25, 535 (1988). [21] A. Le Bail, C. Jacoboni, M. Leblanc, R. De Pape, H. Duroy and J. L. Fourquet, Crystal structure of the metastable form of aluminum trifluoride -AlF3 and the gallium and indium homologues, J. Solid State Chem., 77, 96 (1988). [22] A. Le Bail, J. L. Fourquet, U. Bentrup t-AlF3: Crystal structure determination from X–ray powder diffraction data. A new MX3 corner-sharing octahedra 3D network, J. Solid State Chem. 199, 151 (1992). [23] N. Herron, D. L. Thorn, R. L. Harlow, G. A. Jones, J. B. Parise, J. A. Fernandez-Baca, T. Vogt, Preparation and structural characterization of two new phases of aluminum trifluoride, Chem. Mater., 7, 75–83 (1995). [24] E. Kemnitz and D. H. Menz, Fluorinated metal fluorides oxides and metal fluorides as heterogeneous catalysts, Prog. Solid State Chem., 26, 97 (1998). [25] T. Krahl and E. Kemnitz, The very strong solid Lewis acids aluminium chlorofluoride (ACF) and bromofluoride (ABF)-Synthesis, structure, and Lewis acidity, J. Fluor. Chem., 127, 663–678 (2006). [26] J. A. Wilkinson, Recent advances in the selective formation of the carbon-fluorine bond, Chem. Rev., 92, 505–519 (1992). [27] C. G. Krespan and V. A. Petrov, The chemistry of highly fluorinated carbocations, Chem. Rev., 96, 3269–3301 (1996). [28] M. J. Molina, F. S. Rowland, Stratospheric sink for chlorofluoromethanes – chlorine atomic catalyzed destruction of ozone, Nature., 249, 810–812 (1974). [29] E. Kemnitz and J. Winfield, in Advanced Inorganic Fluorides: Synthesis, Characterization and Applications, A. Tressaud (Ed.), pp. 367–401, Elsevier Science S.A., Amsterdam, 2000. [30] A. Hess and E. Kemnitz, Characterization of catalytically active sites on aluminum oxides, hydroxyfluorides, and fluorides in correlation with their catalytic behavior, J. Catalysis, 149, 449–457 (1994). [31] N. Herron and W. E. Farneth, The design and synthesis of heterogeneous catalyst systems, Adv. Mater., 12, 959–968 (1996). [32] J. L. Delattre, P. J. Chupas, C. P. Grey, A. M. Stacy, Plasma-fluorination synthesis of high surface area aluminum trifluoride from a zeolite precursor, J. Am. Chem. Soc., 123, 5364–5365 (2001). [33] E. Kemnitz, U. Groß, S. Rudiger, C. S. Shekar, Amorphous metal fluorides with extraordinary high surface areas, Angew. Chem. Int. Ed., 42, 4251 (2003). [34] EKLY Hajime, J. L. Delattre and A. M. Stacy, Temperature-dependent halogen-exchange activity studies of zeolite-derived aluminum trifluoride, Chem Mater., 19, 894–902 (2007). [35] A. Wander, B. G. Searle, C. L. Bailey, and N. M. Harrison, Composition and structure of the alpha AlF3 (0001) surface, J. Phys. Chem. B., 109, 22935 (2005). [36] A. Wander, C. L. Bailey, B. G. Searle, S. Mukhopadhyay and N. M. Harrison, Identification of possible Lewis acid sites on the beta-AlF3(100) surface: an ab initio total energy study, Phys. Chem. Chem. Phys., 7, 3989–3993 (2005). [37] A. Wander, C. L. Bailey, S. Mukhopadhyay, B. G. Searle and N. M. Harrison, Ab initio studies of aluminium fluoride surfaces, J. Mater. Chem., 16, 1906–1910 (2006). [38] S. Chaudhuri, P. Chupas, B. J. Morgan, P. Madden and C. P. Grey. An atomistic MD simulation and pair-distribution-function study of disorder and reactivity of alpha-AlF3 nanoparticles, Phys. Chem. Chem. Phys., 8, 5045–5055 (2006). [39] L. Francke E. Durand, A. Demourgues, A. Vimont, M. Daturi and A. Tressaud. Synthesis and characterization of Al3þ, Cr3þ, Fe3þ and Ga3þ hydroxyfluorides: correlations between structural features, thermal stability and acidic properties, J. Mater. Chem., 13, 2330–2340 (2003). [40] M. Moreno, A. Rosas, J. Alcaraz, M. Hernandez, S. Toppi, P. Da Costa, Identification of the active acid sites of fluorinated alumina catalysts dedicated to n-butene/isobutane alkylation, App. Catalysis A: General, 251, 369 (2003). [41] D. Dambournet, A. Demourgues, C. Martineau, S. Pechev, J. Lhoste, J. Majimel, A. Vimont, J.-C. Lavalley, C. Legein, J-Y Buzare´, F. Fayon and A. Tressaud, Nano-structured aluminium hydroxyfluorides derived from -AlF3, Chem. Mater., 20, 1459 (2008).

Microwave-Assisted Route Towards Fluorinated Nanomaterials

67

[42] D. Dambournet, A. Demourgues, C. Martineau, E. Durand, J. Majimel, A. Vimont, H. Leclerc, J.-C. Lavalley, M. Daturi, C. Legein, J.-Y. Buzare´, F. Fayon and A. Tressaud, Structural investigations and acidic properties of high surface area pyrochlore aluminium hydroxyfluoride, J. Mater. Chem., 18, 2421 (2008). [43] D. Dambournet, A. Demourgues, C. Martineau, E. Durand, J. Majimel, C. Legein, J.-Y. Buzare´, A. Vimont, H. Leclerc and A. Tressaud, Microwave synthesis of an aluminum fluoride hydrate with cationic vacancies: structure, thermal stability and acidic properties, Chem. Mater., 20, 7095 (2008). [44] K. J. Rao, K. Mahesh, S. Kumar, A strategic approach for preparation of oxide nanomaterials, Bull. Mater. Sci., 28, 19 (2005). [45] I. V. Kubrakova, R. Khamizov, Fast determination of reaction kinetic parameters with the use of microwave heating. Kinetics of decomposition of organic substances with nitric acid, Russ. Chem. Bull., 54, 1413–1417 (2005). [46] D. S. Jacob, L. Bitton, J. Grinblat, I. Felner, Y. Koltypin and A. Gedanken, Are ionic liquids really a boon for the synthesis of inorganic materials? A general method for the fabrication of nanosized metal fluorides, Chem. Mater., 18, 3162–3168 (2006). [47] D. Dambournet, G. Eltanamy, A. Vimont, J. C. Lavalley, J. M. Goupil, A. Demourgues, E. Durand, J. Majimel, S. Rudiger, E. Kemnitz, J. M. Winfield and A. Tressaud, Coupling sol-gel synthesis and microwave-assisted techniques: A new route from amorphous to crystalline high-surface-area aluminium fluoride, Chem. Eur. J., 14, 6205–6212 (2008). [48] S. Rudiger, G. Eltanany, U. Groß and E. Kemnitz, Real sol-gel synthesis of catalytically active aluminium fluoride, J. Sol-Gel Sci. Techn., 41, 299 (2007). [49] S. Ruediger, U. Groß, M. Feist, H. A. Prescott, S. Chandra Shekar, S. I. Troyanov and E. Kemnitz, Non-aqueous synthesis of high surface area aluminium fluoride – a mechanistic investigation, J. Mater. Chem., 15, 588 (2005). [50] J. Rodriguez-Carvajal and T. Roisnel, Line Broadening Analysis Using Fullprof: Determination of Microstructural Properties, EPDIC8 Uppsala. 2002. [51] P. Thompson, D. E. Cox and J. B. Hastings, Rietveld refinement of Debye-Scherrer synchrotron X-ray data from Al2O3, J. Appl. Crystallogr., 20, 79 (1987). [52] A. Wander, C. L. Bailey, S. Mukhopadhyay, B. G. Searle, and N. M. Harrison, Steps, Microfacets, and Crystal Morphology: An ab Initio Study of -AlF3 Surfaces, J. Phys. Chem. C. 112, 6515–6519 (2008). [53] C. L. Bailey, A. Wander, S. Mukhopadhyay, B. G. Searle, and N. M. Harrison, Characterization of Lewis acid sites on the (100) surface of beta-AlF3: Ab initio calculations of NH3 adsorption, J Chem Phys., 22, 224703 (2008). [54] C.L. Bailey, A. Wander, S. Mukhopadhyay, B.G. Searle and N. M. Harrison, Adsorption of HF and HCl on the beta-AlF3 (100) surface, Phys. Chem. Chem. Phys., 10, 2918–2924 (2008). (See also Chapter 6 in this book.). [55] P. J. Chupas, D. R. Corbin, V. N. M. Rao, J. C. Hanson and C. P. Grey, A combined solid-state NMR and diffraction study of the structures and acidity of fluorinated aluminas: implications for catalysis, J. Phys. Chem. B, 107, 8327–8336 (2003). [56] P. J. Chupas, M. F. Ciraolo, J. C. Hanson and C. P. Grey. In situ X-ray diffraction and solid-state NMR study of the fluorination of g-Al2O3 with HCF2Cl, J. Am. Chem. Soc., 123, 1694–1702 (2001). [57] O. Bose, E. Kemnitz, A. Lippitz and W. E. S. Unger, C 1 s and Au 4 f(7/2) referenced XPS binding energy data obtained with different aluminium oxides, -hydroxides and -fluorides, Fres. J. of Ana. Chem., 358, 175–179 (1997). [58] O. Boese, W. E. S. Unger, E. Kemnitz and S. L. M. Schroeder, Active sites on an oxide catalyst for F/Cl-exchange reactions: X-ray spectroscopy of fluorinated -Al2O3, Phys. Chem. Chem. Phys., 4, 2824–2832 (2002). [59] D. Dambournet, H. Leclerc, A. Vimont, J. C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, The use of multiple probe molecules for the study of the acid-base properties of aluminium hydroxyfluoride having the hexagonal tungsten bronze structure: FTIR and [36Cl] radiotracer studies, Phys. Chem. Chem. Phys., 11, 1369–1379 (2009). [60] R. Chandross, The structure of a new phase of aluminum trifluoride monohydrate, Acta Cryst. 17, 1477 (1964). (It should be noted that the synthesis of a cubic form of an aluminium fluoride monohydrate proposed by Chandross could not be satisfactorily reproduced.).

68

Functionalized Inorganic Fluorides

[61] E. Kemnitz, U. Groß, St. Ru¨diger, G. Scholz, D. Heidemann, S.I. Troyanov, I.V. Morosov and M.-H. Leme´e-Cailleauc, Comparative structural investigation of aluminium fluoride solvates, Solid State Sci., 12, 1443–1452 (2006). [62] L. F. Scatena, M. G. Brown and G. L. Richmond, Water at hydrophobic surfaces: weak hydrogen bonding and strong orientation effects, Science, 292, 908–910 (2001). [63] K. S. Vorres and F. B. Dutton, The fluorides of titanium: X-ray powder data and some other observations, J. Am. Chem. Soc., 77, 2019 (1955). [64] K. Vorres and J. Donohue, The structure of titanium oxydifluoride, Acta Cryst., 8, 25–26 (1955). [65] A. Demourgues, N. Penin, E. Durand, F. Weill, D. Dambournet, N. Viadere and A. Tressaud, New titanium hydroxyfluoride Ti0.75(OH)1.5F1.5 as UV absorber, Chem. Mater., 21, 1275–1283 (2009). [66] Y. Xie, S. Yin, H. Yamane, T. Hashimoto, H. Machida and T. Sato, Microwave assisted solvothermal synthesis of a new compound, pyrochlore-type Sn1.24Ti1.94O3 66(OH)1.50F1.42, Chem. Mater., 20, 493–495. (2008). [67] G. Ferey, Hybrid porous solids: past, present, future, Chem Soc Rev., 37 191–214 (2008). [68] H. Jhung, J.-H. Lee, J. W. Yoon, C. Serre, G. Fe´rey and J.-S. Chang. Microwave synthesis of chromium terephthalate MIL-101 and its benzene sorption ability, Adv. Mater., 19, 121–124 (2007). [69] A. Demourgues, L. Sronek and N. Penin, New Nanostructured Flurocompounds as UV Absorbers, this book, Chapter 8, pp. 229–273, John Wiley & Sons (2010). [70] N. Penin, N. Viadere, D. Dambournet, A. Tressaud, and A. Demourgues, Tuned optical band gap for titanium-based oxy(hydroxyl)fluorides, Proceedings MRS Fall Meeting, Boston, USA, November 2005. [71] N. Penin, N. Viadere, D. Dambournet, A. Demourgues, and A. Tressaud, Synthesis and characterization of Ti-based oxy-hydroxy-fluorides, Proceedings 18th International Symposium on Fluorine Chemistry, Bremen, Germany, July 2006.

3 High Surface Area Metal Fluorides as Catalysts Erhard Kemnitz and Stephan Ru¨diger Humboldt-Universita¨t zu Berlin, Institut fu¨r Chemie, Brook – Taylor – Str. 2, D – 12489 Berlin, Germany

3.1

Introduction

Metal fluorides prepared via the sol-gel fluorination route are nanoscopic and consequently have a high surface area. Metal ions at the surface of a particle cannot have a coordination that is as symmetric as that of ions inside the particle. They are coordinatively undersaturated. That effect together with the strong electron-withdrawing power of fluorine brings about the inherent Lewis acidity of the respective metal ions. Thus, high surface-area aluminium fluoride, HS-AlF3, is an exceptionally strong Lewis acid, as shown in Chapter 1; its acidity corresponds to its position in the Lewis acidity scale calculated for isolated molecules, where AuF5 and SbF5 are the strongest Lewis acids followed by AlCl3 > AlCl2F > AlClF2 > AlF3 although all these aluminium halides exhibit nearly the same acidity [1]. It is interesting to note that magnesium fluoride, typically regarded as neutral compound, also shows distinct Lewis acidity, although not as strong by far as that of HS-AlF3, when prepared via the sol-gel fluorination route [2]. The fact that the acidity of HS-MgF2 diminishes upon heating, as follows from the respective ammonia TPD investigations (Figure 3.1) proves the dependence of the Lewis acidity on the low structural order of the material. Chapter 1 describes how, on one hand, the Lewis acidity of high surface metal fluorides can be adjusted

Functionalized Inorganic Fluorides: Synthesis, Characterization & Properties of Nanostructured Solids  2010 John Wiley & Sons, Ltd

Edited by Alain Tressaud

70

Functionalized Inorganic Fluorides

and how, on the other hand, they can be modified to alter or add other properties than Lewis acidity. Most of these variations have been tested for or have even been used with the aim of improving catalytic properties for selected reactions with often surprisingly good results. These exciting catalytic properties of nanoscopic metal fluorides obtained via fluorolytic sol gel synthesis will be presented in the following. Based on their acidity, often expressed in terms of catalytic activity, the catalyst in the following will be qualitatively differentiated as strong, middle strong, weak, etc. Unfortunately, there is no objective quantitative measure for the strength of acidic surface sites that could be used to rank solid acids as it is well established for acids in aqueous systems by the KS values. On the other hand, FTIR-spectra of probe molecules, e.g. pyridine or CO, adsorbed on acidic surface sites can be used as a measure based on the shift of the absorption peak. The most commonly used probe molecule for routine measurements is pyridine. However, since this is a strong base, its shift of the absorption peak in the FTIR spectrum is too small to give a serious quantification of the acidic strength of a surface site. It is nevertheless an excellent probe molecule to differentiate between the kind of acidity, 32

pretreated at 773 K

30

pretreated at 573 K pretreated at 393 K

28 26 24 22

Intensity/a.u.

20 18 16 14 12 10 8 6 4 2 0 –2 300

400

500

600

700

800

800

Temperature/K Figure 3.1 NH3-TPD of HS-MgF2 preheated at different temperatures. (Reprinted with permission from [2] Copyright (2006) Elsevier Ltd.)

High Surface Area Metal Fluorides as Catalysts

71

that is Brønsted or Lewis, which will be detailed at the appropriate places of this chapter. Carbon monoxide, on the other hand, is a very sensitive probe molecule in terms of peak shift depending on the strength of the acidic surface site. However, since CO is a very weak basic molecule, it is less sensitive regarding weak acidic sites. Alternatively, temperature programmed desorption (TPD) of ammonia gives a comprehensive figure about the distribution of acid strength of surface sites. The principle of this method is, in brief, as follows. The respective solid sample is deposited in a suitable flow reactor, which can be controlled and heated and which is connected with an FTIR-cell. On the precalcined sample (usually between 523 and 573 K) the first dry ammonia is adsorbed at e.g. 373 K. After saturation with ammonia the sample is then flashed by an inert gas stream passing through the connected FTIR cell until ammonia that is no longer desorbed can be detected. Now the reactor will be subjected to a programmed temperature raise and the desorbed ammonia is followed as a function of temperature. Thus a concentration versus temperature profile will be obtained (see for example Figure 3.1) that gives a reflection of the acid site distribution in which the strength of sites is indicated by the temperature (the higher the temperature the stronger the sites) and the concentration of sites of a certain strength is proportional to the intensity of the IR signal. By absorbing the desorbed ammonia by an aqueous acid, the overall amount of desorbed ammonia can be exactly determined that corresponds to the square below the desorption graph. Based on this one may go to determine different ranges of acidic strengths. This might be acceptable as long as different solids that are similar in nature (e.g. metal fluorides) are compared and as long as the same equipment is used. However, since these profiles depend on several parameters (surface morphology, design of the desorption cell, heating rate, detection system etc.) such data would not really allow any serious comparison of different samples measured in different labs. Hence, although both of the techniques mentioned above are very often used – and are also used by the authors – it seems less useful to apply real quantification because of the problems briefly mentioned. If not otherwise stated the different catalytic parameters used in this chapter have the following meaning: Conversion defines the part of converted compound A (n0AnA) in relation to the starting concentration (n0A); X ¼ (n0AnA)/n0A Selectivity defines the part of a desired product P among all the products formed: SP ¼ ðnP  veduct Þ=nP  ðn0educt  neduct Þ Yield defines the part of the desired product in the reaction mixture, YP ¼ SPXA

3.2

High Surface Area Aluminium Fluoride as Catalyst

The very high Lewis acidity of HS-AlF3 causes it to react easily with most electron pair donating solvents such as alcohols or aqueous ones and many reactants whereby the acidic sites of HS-AlF3 become blocked. Therefore, if its high Lewis acidity is of interest for catalytic reactions it can be used at best in fluoro-organic reactions although a few others have also been reported. Seven reactions of fluoro-organic compounds have been successfully catalysed by HS-AlF3 (Equations (3.1) to (3.7)) [3], the use of the reactions in

72

Functionalized Inorganic Fluorides

Equations (3.1) to (3.4) as test reactions to assess the Lewis acidity of the catalyst is already described in Chapter 1. 5 CCl2 F2 ! CCl3 F þ 3 CCl3 F þ CCl4

(3:1)

3 CHClF2 ! CHCl3 þ 2 CHF3

(3:2)

CBrF2 CBrFCF3 ! CF3 CBr2 CF3

(3:3)

CCl2 FCClF2 ! CCl3 CF3

(3:4)

F3 C-CHF-CHF2 ! F3 C-CF¼CHF þ HF

(3:5)

F3 C-CH2 -CF3 ! F3 C-CH¼CF2 þ HF

(3:6)

F3 C-CH2 F ! F2 C¼CHF þ HF

(3:7)

The dismutation reactions (3.1) and (3.2) proceed only under the catalytic action of a Lewis acid albeit a moderate Lewis acidity is sufficient. Thus, HS-MgF2 is reported to catalyse the dismutation reaction (Equation (3.2)) in a flow reactor with 3% conversion at 573 K and 60% conversion at 623 K whereas common MgF2 is not at all active [2]. With the much stronger Lewis acid HS-AlF3 reactions (Equation (3.1)) and (Equation (3.2)) proceed with almost 100% conversion even at room temperature [4]. The isomerization reaction (Equation (3.3)) proceeds only under the catalytic action of the strongest Lewis acids known, SbF5 (at elevated temperatures, 383 K), aluminium chlorofluoride (ACF), and HS-AlF3 [5]. As with ACF, with HS-AlF3 almost 100% conversion was observed at room temperature in a batch reaction. The isomerization reaction (Equation (3.4)) also proceeds very effectively in the presence of HS-AlF3. In a detailed study comparing different Lewis acidic Al-F and Al-Cl based catalysts, which is described in Chapter 1, HS-AlF3 proved to be almost equal to ACF and superior to the AlCl3 and -AlF3, which are both not active at room temperature. However, -AlF3 is well known as a good and selective catalyst for this reaction at temperatures above 553 K [6–8]. AlCl3 however, became very active after being converted into ACF by a chemical reaction with CCl2FCClF2 [5]. Hence, the latter experiments proved that solid HS-AlF3 is a better Lewis-acid catalyst than solid AlCl3. The well known excellent Lewis acidity of AlCl3, – and based on this its high catalytic activity in many organic reactions for which in contrast the currently known AlF3 phases are totally inactive – is just a result of the better solubility of AlCl3 in comparison to the practically insoluble AlF3. Thus, the catalytic difference between these two crystalline phases is exclusively caused by their different solubility in organic solvents. Although HF behaves towards HS-AlF3 as base adsorbed at and blocking the strongest Lewis acid sites [7], HS-AlF3 has also proved to be an effective catalyst for dehydrofluorination reactions (Equations (3.5) to (3.7) [3]. It is known that adsorbed HF is almost completely released at above 573 K as can be seen in Figure 3.2 [7]. Therefore, such reactions have to be carried out at temperatures distinctly higher than those needed for the reactions in Equations (3.1) to (3.4). At 623 K 1,1,1,2,3,3-hexafluoropropane is dehydrofluorinated with 95% conversion to the 1,1,1,2,3-pentafluoropropene with 88% selectivity for the cis-isomer and 12 to the trans-isomer (Equation (3.5)). No formation of any 1,1,1,3,3-pentafluoropropene was

High Surface Area Metal Fluorides as Catalysts DTG /(% min–1) DTA /μV Ion Current ·10–10/A

TG (%) 100.0 TG

85.0

DTA ↑ exo

10

818

95.0 DTG 90.0

73

0 –14.43 –10

m19(×40)

–20

80.0

–30

806

75.0

–40

70.0

m18

65.0

m20(×40)

–50 –60

60.0 373

473

573

673 773 Temperature/K

0 –0.10 –0.20

8.0 7.0 6.0 5.0

–0.30

4.0

–0.40

3.0

–0.50

2.0

–0.60

1.0

–0.70

0

873

Figure 3.2 TA/MS curves of HS-AlF3 loaded previously with HF, and IC curves for m/z 18 (H2Oþ), m/z 19 (Fþ), and m/z 20 (HFþ). The identical curve shapes for m/z 19 and 20 can be seen as proof that they originate from the same molecule, i.e. HF, whereas m/z 18 (H2Oþ) is different. (Reprinted with permission from [7] Copyright (2005) Royal Society of Chemistry.)

observed under these conditions. Under similar conditions at 673 K 1,1,1,3,3,3-hexafluoropropane reacts to 1,1,1,3,3-pentafluoropropene with 33.5% conversion and 66% selectivity towards 1,1,1,3,3-pentafluoropropene (Equation (3.6)). With decreasing temperature in both reactions the conversion decreases but the selectivity increases. Dehydrofluorination of 1,1,1,2-tetrafluoroethane to trifluoroethene (Equation (3.7)) also proceeds at 623 K with 20% conversion and 78% selectivity toward the desired product. The reactions in Equations (3.6) and (3.7) necessitate an attack on the respective CF3 group, recognized to be very stable. The activation of these groups within the progression of the reactions gives evidence for the unusually high Lewis acidity of the catalyst although this was to a minor extent also observed at fluorinated crystalline aluminium fluorides and fluorinated aluminas [8]. Other reactions reportedly catalysed by HS-AlF3 are shown in Equations (3.8) to (3.10).

(3.8)

+ H H

H

H

+

H H

(3.9)

ðCH3 Þ 3 CCl ! CH2 ¼CðCH3 Þ2 þ HCl; CH2 ¼ CðCH3 Þ2 ! CH3 CH ¼ CHCH3 (3:10) Isomerization reactions (3.8) and (3.9) have been carried out at 573 K or 473 K, respectively, in a flow reactor, whereby in reaction (3.8) 83% conversion and in reaction (3.9) 47% conversion was observed [3]. Reaction (3.10), dehydrochlorination of tBuCl

74

Functionalized Inorganic Fluorides

followed by isomerization of the 2-methylpropene to butene-(2), was observed at room temperature in course of radiotracer experiments and is described in Chapter 4. The before mentioned isomerisation reaction of 2-methylpropene to butene-(2) was observed for the first time under such circumstances indicating the very strong Lewis acidity of HS-AlF3. A solid catalyst designed for use in a continuous-flow reactor has to meet some physical requirements, such as mechanical stability and low flow resistance, requirements that are not a priori fulfilled by HS-AlF3, which is a fine powder. Fortunately, there is a convenient way to support HS-AlF3 on g-alumina (see Chapter 1), and the supported catalyst showed similar catalytic activity in all the above reactions [10].

3.3

Host-Guest Metal Fluoride Systems

According to the Tanabe model [11], originally developed for metal oxide host-guest systems but later adapted for metal fluoride host-guest systems [12], the kind of acidity (Lewis or Brønsted) can be predicted based on the anion coordination of the host and guest metal ions, respectively. This is a pure geometric model but allows with up to 90% accuracy to predict whether Brønsted or Lewis acidity will be generated by creating solid solutions. However, it does not give any prediction of the strength of acidity. In brief, the following assumptions are made according to this approach applied for a solid solution two different metal fluorides (MFx and M’Fy). (i) The coordination number of both of the metal cations in MFx and M’Fy are maintained even when mixed. (ii) The coordination number of the fluoride ion of the major (host) metal fluoride component is retained for all the fluorides in a ternary metal fluoride system. Suppose, for example, that CrF3 is introduced as guest in the MgF2-host system. The coordination numbers of both metals are 6 and remain even in the guest-host system, whereas the coordination number of all the fluoride ions is 3. Keep in mind that the CN of F in CrF3 is just 2! If we now count the electrostatic bonding power, we have the following situation: the positive charge of Mg is 2þ and will be distributed to six surrounding (coordinating) F-atoms, i.e. þ2/6 ¼ þ1/3 charge is directed to each F-atom. On the other hand, the charge of F is distributed towards three coordination Mg atoms, e.g. –1/3. This means that the charges are perfectly balanced between Mg and F as expected. However, the situation is different in case of the host cation Cr3þ: since Cr3þ is coordinated by 6 F atoms, the charge is distributed by þ3/6 ¼ þ1/2 to each surrounding F atom. However, since all the F atoms exhibit the same coordination number 3 the charge coming from the F atom is just –1/3. For each Cr3þ cation the overall charge balance can be calculated by 6x (þ1/2 – 1/3) ¼ þ1. The conclusion from this is that by introducing CrF3 into a MgF2 host a local positive charge at the chromium site is generated acting as a Lewis acidic site. In an opposite way (introduction of MgF2 into a CrF3 host), negative charges will be generated which becomes compensated by attraction of protons thus creating Brønsted sites (see details [12]). Mixed metal fluoride systems where the cations of the minor component can principally substitute cations of the major component in the lattice therefore open an easy way to

High Surface Area Metal Fluorides as Catalysts

75

Table 3.1 Benzoylation/acetylation of anisole, i-propylation of benzene and dismutation of CCl2F2 over host/guest catalysts (Data taken from [13] by permission of Elsevier Publishers) Catalyst

Fe/MgF2 Ga/MgF2 In/MgF2 V/MgF2 a b c

Benzoylation with Ph-COCl

Acetylation with AcOAc

Xb(%)

Yc(%) 4-MBPh

X (%)

100 100 100 0

69 77 79 0

45 15 nda 9

Y (%) 4-MAPh 34 6.5 – 1

i-propylation with CH3CHClCH3

Dismutation of CCl2F2

X Y (%)-cumene (%)

Temp. (K)

X (%)

100 nd nd nd

623 623 623 623/573 523/473

25 30 5 85/56/ 28/10

84 – – –

nd: not determined. X: conversion. Y: yield.

create metal fluoride-based solid catalysts with adjusted Lewis acidity (see Chapter 1). The necessity for adjusted Lewis acidity follows from a comparison of HS-AlF3 and -AlF3 as solid catalysts. Although HS-AlF3, as very strong Lewis acid, should be an ideal catalyst for all Lewis acid-catalysed reactions, it cannot be used in many of these reactions because it becomes almost immediately deactivated by often irreversible adsorption of educts and/ or necessary solvents thus blocking the Lewis acid sites. -AlF3 on the other hand is not a strong enough Lewis acid for many reactions. Many MgF2/MF3 systems are reported with HS-MgF2 being the major component, the host, and the guest metal fluoride amounts to up to 30%. Such systems, which should be rather considered as solid solutions and which might be doped systems, have been reported with Fe3þ, Ga3þ, V3þ, In3þ, and Cr3þ as guest cations [13–15]. For MgF2 the precursor was always an alkoxide, whereas fortunately, for the guest component a broader range of compounds could be used as precursor. It was found that surface area and pore size as well as Lewis acidity and catalytic activity depend on the nature of the guest cation and its concentration, as expected, and also partly on the chemical nature of the compound that was used to introduce the guest metal during the sol-gel synthesis. The effect of the respective guest cation follows from Table 3.1. All these catalysts were prepared from Mg(OMe)2, MCl3, and In(OiPr)3 (85:15), respectively [13]. The catalysts showed very different activity for dismutation of CCl2F2 (Equation (3.1)) despite only moderate variations in surface area (100155 m2/g), and the best one, V/MgF2, which is clearly inferior to HS-AlF3, was on the other hand not at all active in the benzoylation reaction. Obviously, the Lewis acidity of V/MgF2 is that high, as proved by its dismutation activity, that at the acid sites of the solid components of the benzoylation reaction adsorb irreversibly thereby blocking the catalyst as known for HS-AlF3. It is, however, worth noting that the simple solid Fe3þ/MgF2 catalyst was rather active for Friedel-Crafts benzoylation (Equation (3.11)) and acylation as well as alkylation (Equation (3.12)). CH3 OC6 H5 þ C6 H5 -CðOÞCl ! CH3 OC6 H4 -CðOÞ-C6 H5 þHCl cat

C6 H6 þ C3 H7 Cl ! C6 H5 -C3 H7 þHCl

(3:11) (3:12)

76

Functionalized Inorganic Fluorides

In a more detailed investigation of the Fe3þ/MgF2 catalyst system (Fe:Mg ¼ 15:85) three types of Fe3þ precursor, i.e. FeCl3, Fe2(SO4)3 and Fe(OCH3)3, and two ways of postfluorination, i.e. with CCl2F2 or with HF, were employed [14]. The surface areas of the HF fluorinated catalysts (174–298 m2/g) were markedly higher than of the CCl2F2 fluorinated ones (75–104 m2/g). The latter showed, however, a somewhat higher catalytic activity for CCl2F2 dismutation (Equation (3.1)) and especially anisole benzoylation (Equation (3.11)) with 97–100% conversion and 60–82% selectivity whereas the HF fluorinated catalysts gave 41–89% conversion and 42–62% selectivity. An Fe(OCH3)3 based and HF postfluorinated catalyst was repeatedly successfully used in benzoylation, giving no evidence of leaching. Two catalysts of the CCl2F2 postfluorination group, based on FeCl3 (named MgFeClR) and Fe2(SO4)3 (named

B 90

553 K

80

593 K

633 K

70 60 50 40 30

Relative content (%)

20 10 0

conv

R123

R124

R125

others

conv

R123

R124

R125

others

A 70 60 50 40 30 20 10 0

Figure 3.3 Hydrofluorination of C2Cl4 over MgFeClR (lower graph, obtained from FeCl3 and postfluorinated by CCl2F2) or over MgFeSR (upper graph, obtained from Fe2(SO4)3 and postfluorinated by CCl2F2). (Reprinted with permission from [14] Copyright (2004) Elsevier Ltd.)

High Surface Area Metal Fluorides as Catalysts

Conversion of CFC12, %

90 80

CrClMg15 CrOMg8

CrAcMg15 CrOMg25

77

CrOMg15

70 60 50 40 30 20 10 0 350

400

450

500 550 Temperature, K

600

650

Figure 3.4 Dismutation of CCl2F2 over CrF3/MgF2 catalysts at tR ¼ 2 s; CCl2F2:N2 ¼ 1:5. (Reproduced from [15] by permission of Elsevier.)

MgFeSR), respectively, have also been tested for hydrofluorination of C2Cl4 (Equation (3.13)). CCl2 CCl2 þðx þ 1ÞHF ! C2 HCl4  x Fxþ1 þxHCl

(3:13)

The technically useful reaction (3.13) cannot be catalysed by HS-AlF3 because of irreversible adsorption of HF. Therefore, it is a very interesting result that with the Fe3þ/ MgF2 catalysts, which are highly Lewis acidic although not as high as HS-AlF3, the hydrofluorination proceeded well (Figure 3.3) [14]. There is a similarly detailed study of the Cr3þ/MgF2 guest/host system scrutinizing physicochemical characteristics and catalytic properties of both, differently prepared and composed solids [15]. The catalysts were prepared from Mg(OMe)2 and CrCl3 (Mg:Cr ¼ 85:15; named CrClMg15), Cr3Ac7(OH)2 (85:15; CrAcMg15)) and CrO3, prior to use reduced by reaction with MeOH, (82:8, 85:15 and 75:25; CrOMg8, CrOMg15 and CrOMg25). CCl2F2 was employed as postfluorination agent. The XRD patterns of all catalysts show very broad reflexes only for MgF2 and give no hint for other phases. Their surface area ranged from 65 m2/g (CrAcMg15) over 99 m2/g (CrOMg25) and 106 m2/g (CrClMg15) to 167 m2/g (CrOMg8) and 175 m2/g (CrOMg15). CrOMg15 had the highest surface area and was the most active catalyst for CCl2F2 dismutation (Figure 3.4), indicating the highest Lewis acidity of all the preparations. Comparing the CrOMg catalysts of different molar Cr:Mg ratio it becomes evident that about 15 mol% guest cations is the optimum composition. This holds also for the dismutation of CHClFCF3 according to Equation (3.14) the results of which are shown in Figure 3.5. 2CHClFCF3 ! CHF2 CF3 þCHCl2 CF3

(3:14)

Obviously, guest cations up to a concentration of about 15 mol% can readily become incorporated into the lattice of the host, provided the two cations are of comparable size, enhancing the Lewis acidity and the derived catalytic activity of the solid solution compared to the neat host system.

78

Functionalized Inorganic Fluorides 100

CHCl2CF3

CHClFCF3

Relative concentration (%)

90 80 70 60 50 40 30 20 10 0

CrClMg15 CrAcMg15 CrOMg15

CrOMg8

CrOMg25

Figure 3.5 Dismutation of CHClFCF3 over Cr/MgF2 catalysts at 573 K; tR ¼ 2 s; CHClFCF3:N2 ¼ 2:1.5. (According to Equation (3.14), the amounts of CHF2CF3 have to correspond with the respective amounts of CHCl2CF3. However, the experimental values for CHF2CF3 are smaller due to losses because of its higher volatility. They are therefore omitted.) (Reprinted with permission from [15] Copyright (2005) Elsevier Ltd.)

3.4

Hydroxy(oxo)fluorides as Bi-acidic Catalysts

Depending on the nature of the metals the M-OH units created in a metal fluoride system may either react as a base or as a (Brønsted) acid. Thus, replacement of F by OH, or vice versa, should result in an alteration of the basicity of M-OH groups, e.g. in MgF2x(OH)x derivatives as compared to Mg(OH)2 and in the same way of the (Brønsted) acidity, e.g. in AlF3x(OH)x phases as compared to Al(OH)3. Thus, some of the fluoride ions of a metal fluoride can be replaced by hydroxyl and/or oxide groups in the course of its preparation by introducing defined amounts of water at different steps of the synthesis, modifying the properties. A number of magnesium fluoride-derived materials have been prepared starting from Mg(OMe)2 and understoichiometric amounts of HF, which immediately form a sol to which twice the theoretically needed amount of water was added. One preparation started from a stoichiometric Mg:HF ratio with supplementary water added to the sol. The final gels were dried and calcined at 350 C [16]. The preparations, named MOF-0, MOF-0.4, MOF-1.2, MOF-1.6, and MOF-2.0 according to relative amount of HF to Mg used, had analytically about 80% of the nominal F content, obviously due to incomplete fluorination in the gel phase, as is known for HS-MFx syntheses. Despite their designations the hydroxyl groups introduced are almost completely dehydrated at the calcination temperature, creating oxidic O atoms. The materials showed graded physico-chemical properties in terms of surface area, ranging from 104 m2/g (MgF2) to 387 m2/g (Mg(OH)1.2F0.8), and in terms of XRD pattern, FTIR spectra, 19F MAS NMR spectra as well as Mg 1s, F 1s and O 1s binding energies [16]. The results of Michael addition experiments are very interesting, i.e. addition of

High Surface Area Metal Fluorides as Catalysts

79

2-methyl-cyclohexane-1,3-dione to methyl vinyl ketone, employing the synthesized magnesium hydroxide fluorides as catalyst compared to MgO and crystalline MgF2 catalysts, shown in Figure 3.6. MgF2 was not at all active (so it is not included in Figure 3.6), whereas MgO was very active giving 70% yield in less than 1 h. However, after 1 h the yield started to decrease to about 20% in 24 h caused, by consecutive reactions of the primary addition product. Likewise, with commercial Mg(OH)2 a maximum yield of 70% is reached after 4 h, followed also by a decline to 45% in 24 h. The F-containing catalysts showed a more differentiated activity. The catalyst with the lowest F content, h-Mg(OH)1.6F0.4, was the

A 100 MOF-0 MOF-0.4 MOF-1.2 MOF-1.6 MOF-2.0 Mg(OH)2

Yield of Michael adduct (%)

90 80 70 60 50 40 30 20 10 0 0

2

4

6

8

20

22

24

Time/h

Selectivity of Michael adduct (%)

B

100 MOF-0 MOF-0.4 MOF-1.2 MOF-1.6 MOF-2.0 Mg(OH)2

90 80 70 60 50 40 30 20 10 0 0

2

4

6

8

20

22

24

Time/h

Figure 3.6 Michael addition of 2-methyl-cyclohexane-1,3-dione to methyl vinyl ketone catalysed by Mg(OH,F)2 catalysts. Yield (above) and selectivity (below) versus reaction time. (Reprinted with permission from [16] Copyright (2005) Elsevier Ltd.)

80

Functionalized Inorganic Fluorides

most active of all, with about 90% yield after 2 h, which, however, decreased rapidly as in case of MgO, indicating that there are basic sites at the catalyst that are too strong, promoting consecutive reactions. The two catalysts with the highest F content, hMg(OH)0.4F1.6 and h-MgF2, showed a totally different activity – they were less active but gave the highest yield after 24 h because of absence of consecutive reactions. Accordingly, they were very selective (Figure 3.6). The most active catalyst h-Mg(OH)0.4F1.6 was also used in three other Michael addition reactions, i.e. addition of 2-acetylcyclopentanone, 2-acetylcyclohexanone, and 2-methoxycarbonyl-cyclopentanone to methyl vinyl ketone. These reactions proceeded also with high (75–90%) yields and 100% selectivity [16]. The great dependence of activity on the F content of the samples was explained as follows. Both, F– and O2– are by definition Lewis bases but pure MgF2 did not give any reaction at all, meaning the F-sites do not act as Lewis basic sites. In contrast the magnesium oxofluorides carrying O2 are very active. Hence, the explanation is that O2 sites at the solid surface are excellent proton acceptors forming Brønsted basicity, which is necessary for the catalytic performance but F–-ions are not at all. The F atoms, on the other hand, reduce the basicity of the O atoms due to their electron-withdrawing effect. Thus, by changing the F/O ratio the basicity of the O atoms can be tuned down to a level sufficient for catalysing the addition reaction but too low for the consecutive reactions. Another closely related synthesis route to partly hydroxylated metal fluorides starts from metal alkoxide in alcohol, which is reacted with the stoichiometric amount of hydrogen fluoride dissolved in different amounts of water thus enabling the competition between fluorolysis and hydrolyis, which to some extent is ruled by the HF to H2O ratio and results in partly hydroxylated fluorides. Thus, Mg(OCH3)2 dissolved in MeOH was reacted with stoichiometric amounts of aqueous HF as 40wt%, 57wt%, 71wt%, and 87wt% solution yielding a viscous, transparent gel, which was dried under vacuum firstly at room temperature and finally at 70 C for 5 h. The hydroxylated MgF2 products, named MgF2-40, MgF2-57, MgF2-71 and MgF2-87 and having surface areas ranging from 180 to 420 m2/g, have been employed as catalyst for (all-rac)-[]-tocopherol (vitamin E, 3) synthesis through condensation of trimethylhydroquinone (TMHQ, 1) with isophytol (IP, 2) (Equation (3.15)), a reaction known to need acid catalysis [17]. HO

HO

+ OH TMHQ, 1

OH IP, 2

C14H29

catalyst –H2O

O

C14H29

(3.15)

3

Table 3.2 shows some results of tocopherol synthesis obtained with different MgF2 based catalysts. Both crystalline MgF2 (entry 1) and HS-MgF2 prepared with very little or no water (entries 7 and 8) were not at all active, even after prolonged reaction time. As crystalline MgF2 exhibits almost no acidity its inactivity was to be expected. On the other hand, a HS-MgF2 catalyst prepared with 71% aqueous HF resulted in total conversion of isophytol and almost 100% selectivity to (all-rac)-[]-tocopherol (entry 6). The activities of the catalysts do not correspond to their respective numbers of acid centres (Table 3.2). Likewise, the 19F MAS NMR spectra do not correspond to the activity.

High Surface Area Metal Fluorides as Catalysts

81

Table 3.2 Influence of the key surface features and the reaction parameters on the catalytic performances of tocopherol synthesis Entry Catalyst 1 2 3 4 5* 6** 7 8

Number of A.C./m2

MgF2-C MgF2-40 MgF2-57 MgF2-71 MgF2-71 MgF2-71 MgF2-87 MgF2-100

n.d. 5.4  6.6  3.6  3.6  3.6  4.8  8.4 

IP/catmolar Time, (min) Conversion of ratio isophytol n.d. 119 76 123 31 123 60 37

1017 1017 1017 1017 1017 1017 1017

1800 300 300 300 60 180 360 360

Yield of tocopherol (%)

0 100 100 100 100 100 0 0

0 76.3 82.6 87.0 92.6 > 99.9 0 0

Reaction conditions: 50mg of catalyst; T ¼ 373 K; TMHQ/IP ¼ 1/1; solvent: heptane/propylene carbonate¼ 50/50 (v/v). Yield is based on IP (C ¼ 100%).* 200mg of catalyst.** TMHQ/IP ¼ 2/1; A.C. ¼ acidic centres calculated from NH3-TPD and N2 sorption isotherms.

BPy

LPy

L.A.

MgF2-100 MgF2-87 MgF2-71 MgF2-57 B.A.

MgF2-40

2000

1800

1600 1400 Wavenumber [cm–1]

1200

Figure 3.7 Infrared spectra of pyridine adsorbed on HS-MgF2 preparations

The resonance signal (–198 ppm) is at almost identical position compared to that of crystalline MgF2, but with increasing water content the signal showed a slight deformation to low field, which can be assigned to increased structural disorder due to increased number of OH groups [18]. The IR spectra of adsorbed pyridine (Figure 3.7) in the region characteristic for adsorption on Brønsted acid sites and on Lewis acid sites are more informative. Thus, MgF2-100 and MgF2-87 exhibit almost only Lewis acid sites, responsible for the shoulder at 1611 cm1. This can be seen in an even more pronounced way from the intensity of the peak at ca. 1450 cm1 which decreases with increasing water content of the HF used for the synthesis. At the same time, the peak at 1545 cm1 increases. Obviously, Brønsted sites are increasingly created at the expense of Lewis sites. With increasing water content, the Lewis acid sites

82

Functionalized Inorganic Fluorides

diminish and Brønsted sites occur, the ratio of Lewis to Brønsted sites is adjusted and ‘biacidic’ material is created. Evaluating the experimental results (Table 3.2) one can see that the predominantly Brønsted acidic MgF2-40 catalyst is highly active (100% conversion) but less selective. On the contrary, the predominantly Lewis acidic MgF2-100 and MgF2-87 catalysts are not active. Obviously, the graded Brønsted acidity of the ‘bi-acidic’ MgF2-71 catalyst provides optimum conditions for the overall reaction, preventing the formation of consecutive products. The observed high regioselectivity to (all-rac)-[]-tocopherol might be due to the structural peculiarities of the catalyst as it is reported that very bulky polymerbound catalysts improve the regioselectivity of this reaction [19]. Comparing the properties of the hydroxylated magnesium fluoride catalysts, a relationship between the extent of hydroxylation and Brønsted acid/base behaviour becomes obvious. Contrary to expectations, Mg-OH bonds can be Brønsted acidic when created to a very small extent in an MgF2 host phase, obviously, due to the electron-withdrawing effect of the fluoride ions. However, the higher the degree of hydroxylation the lower the Brønsted acidity or in other terms the higher the Brønsted basicity. Very small amounts of hydroxyl groups are strong Brønsted acids as proved for tocopherol synthesis, and very high numbers of hydroxyl groups corresponding with smaller amounts of fluoride are Brønsted bases as proved for the Michael addition reactions. A simple test with aqueous acid/base indicator showed exactly the same relation. [18]. In the light of the findings with hydroxylated MgF2 catalysts, HS-AlF3 with its exceptional high Lewis acidity but small Brønsted acidity should be a poor catalyst for tocopherol synthesis. In fact, after as long as 1200 min reaction time (conditions as given in Table 3.2) only 70% IP conversion were detected, but the reaction stopped at an intermediate stage at phytylhydroquinone, PHQ. The transformation of PHQ to (all-rac)-[]-tocopherol could not be provoked by the strong Lewis and weak Brønsted acid HS-AlF3 [20]. However, with a partly hydroxylated catalyst (‘AlF3-H’) prepared from aluminium isopropylate reacted in propanol with stoichiometric amounts of HF added as 50% (w/w) solution in water and followed by drying the gel formed under vacuum (finally at 70 C for 5 h) the conversion of IP was completed within 1 h and (all-rac)-[]-tocopherol formed in very high yield. Obviously, it is not strong Lewis acid sites but moderate to medium-strong Brønsted acid sites that are needed to catalyse the overall reaction. The difference in the acidities of HSAlF3 and AlF3-H follows from the respective IR spectra of adsorbed carbon monoxide gas shown in Figure 3.8. Compared to neat HS-AlF3 AlF3-H does not have super-strong Lewis acid sites indicated by the almost complete absence of CO IR absorption bands in the region above 2200 cm1, which is indicative for very strong Lewis acid sites. The strongest absorption occurs in the region between 2160 and 2180 cm1, the intensive band at 2172 cm1 can be assigned to Brønsted acid sites, proving the dominance of Brønsted acidity. This is supported by the spectra of adsorbed lutidine in Figure 3.9 In the spectra there are two bands at 1655 and 1631 cm1 corresponding to protonated lutidine, observable only with moderate to strong Brønsted acid sites at the solid, and in case of high lutidine partial pressure (Figure 3.9a) there are additional absorption bands at 1581, 1595 and 1604 cm1 corresponding to physisorbed and weakly coordinated species, respectively. The results obtained with AlF3-based catalysts confirm the findings with hydroxylated MgF2 catalysts. They show that introducing a graded hydroxylation upon the sol-gel fluorination synthesis opens possibilities for fine-tuning the strength and balance of Lewis/Brønsted acidity and thus extends substantially the fields of application of metal fluoride-based catalysts.

High Surface Area Metal Fluorides as Catalysts

83

Adsorption

2220

2200

2160

2180

2140

Wavenumbers /cm–1

Figure 3.8 Infrared spectra of increased doses of CO adsorbed on AlF3-H at 100 K. (Reprinted with permission from [20] Copyright (2008) Wiley-VCH Verlag GmbH.)

1581 1655

1595 1631

1604

a b c d

1700

1650

1600

1550

–1

Wavenumbers/cm

Figure 3.9 Infrared difference spectra after adsorption of lutidine on AlF3-H activated at 373 K. (a) equilibrium pressure at RT, (b) same as (a) but after evacuation at RT, (c) evacuated at 323 K, (d) evacuated at 373 K. (Reprinted with permission from [20] Copyright (2008) WileyVCH Verlag GmbH.)

Since it seems obvious that these new catalysts might be applied for other related reactions, we also tested aluminium and magnesium hydroxyfluorides for the cyclization of citronellal yielding (–)–isopolegol [21]. Some of these catalysts yielded extremely high chemo- and diasterioselectivity for (–)–isopolegol, which has not been achieved so far with other catalysts –not even with homogenous ones.

84

3.5

Functionalized Inorganic Fluorides

Oxidation Catalysis

Typically, oxidation catalysts are oxides of transition metals, the oxidation state of which changes easily. Vanadium oxide species (VOx) are well-known examples for oxidation catalysts. Their performance depends on the dispersion and coordination of VOx and on the surface properties, especially the acidity of the support on which VOx is deposited [22–26]. Thus, VOx supported by materials such as ZrO2 and TiO2, which have strong Brønsted acid sites on their surface, is very active in propane oxidation, however, its selectivity to propene, the product that is aimed for, is very low at higher conversion degrees and CO/CO2 formation dominates besides H2O because of formation of very reactive carbocations as intermediate. On the other hand, without acid sites, as in case of VOx/SiO2, very low conversion of propane occurs, obviously because the propane becomes inactive [27]. Although it is still a matter of debate in literature it is obvious that acid surface sites, beside oxygen mobility in the catalyst, are a major source of COx-formation. Surface Brønsted acid sites, in particular, are considered to activate the propane molecule in such a way that intermediate carbocations are formed which are too reactive and, hence, proceed to form COx. With the access of highly Lewis acidic HS-AlF3 it seemed plausible to use this as support for the oxygenating component, VOx because Lewis acidic supports and supports with balanced Lewis/Brønsted acidy should be able to activate the propane without formation of carbocations with their excessive reactivity, giving catalysts with improved performance. The expectation was to prevent the formation of reactive carbocationic species, which should significantly improve the selectivity toward propene and should suppress COx-formation. The sol-gel synthesis of metal fluorides provides a convenient access to VOx catalysts supported by high surface-area metal fluorides exhibiting Lewis acid and also some Brønsted acid sites (see Chapter 1). Aluminium fluoride-based VOx catalysts have been prepared from aluminium triisopropylate and vanadium(v) oxytripropoxide with both stoichiometric and understoichiometric amounts of HF/Al. In some experiments was water also added. This way, after work up and calcination, VOx catalysts have been obtained that are supported by high surface aluminium fluoride or aluminium oxyfluoride. The performance of a series of VOx/Al oxyfluoride catalysts (prepared with Al:HF ¼ 1:2) for oxidative dehydrogenation (ODH) of propane is presented in Figure 3.10. Although the yield of propene does not exceed about 12% at best, the performance of the catalysts is remarkable insofar as, for the first time, the yield of propene is higher than that of COx, with the 10% vanadium catalyst. The good performance of the aluminium fluoridesupported VOx catalysts is due to the nature of the support and not a result of the sol-gel preparation route as follows from a direct comparison of VOx/aluminium fluoride (‘VAlF’) and VOx/aluminium oxide (‘VAlO’) catalysts prepared similarly via the sol-gel route. In Figure 3.11 compares the performance of VAlF and VAlO catalysts with varied V content for propane-ODH. From Figure 3.11 (a) follows that with aluminium oxide-based catalysts high conversion rates have been obtained, even with only 10 mol% V content, however, predominantly as undesired total combustion with a low yield of propene. A rather different picture was obtained with the aluminium fluoride-based catalysts (Figure 3.11 (b)). Catalysts with low V content still gave acceptable conversion rates but the conversion resulted predominantly in the aimed propene. Only with higher V content did the conversion rate increase to nearly the same values as in the case of VAlO catalysts, however, the yield

High Surface Area Metal Fluorides as Catalysts

85

Conversion X & yield Y (%)

40 X (C3H8) Y (C3H6) Y (CO) Y (CO2)

30

20

10

0 0

5

10 15 20 V-content/mol-%

25

30

Figure 3.10 Oxidative conversion of C3H8 (X) and yields (Y) of C3H6, CO and CO2 over VOx containing aluminium oxyfluorides with different V content. (Reprinted with permission from [28] Copyright (2005) Royal Society of Chemistry.)

of propene did not increase accordingly. The superior performance of the VAlF-type catalysts led to their thorough investigation, mostly in comparison with VAlO catalysts [29]. Thus, VAlF (from Al:HF ¼ 1:3) and VAlO have, in the catalytically interesting range of up to 20 mol% V, rather similar specific surface area whereas pore volume and especially the number of acid sites (from NH3-TPD) are significantly higher for VAlF. Infrared spectroscopic investigations of adsorbed pyridine showed that the VAlF acid sites are predominantly Lewis sites. X-ray diffraction measurements showed that VAlF is amorphous and the diffraction patterns of VAlO contain only very broad reflexes of Al-O-species. The absence of VOx related diffraction patterns prove that they are incorporated in the respective Al matrices. The 27Al MAS NMR and 19F MAS NMR spectra of VAlF, depicted in Figure 3.12, resemble very closely those of HS-AlF3. The chemical shift of dF ¼ –167 ppm and the large spinning side bands are indicative of an Al in a highly distorted AlF6 environment [29]. The 51V MAS NMR spectra (Figure 3.13) and ESR spectra (Figure 3.14) of the VOxdoped aluminium fluoride and oxyfluoride systems revealed that vanadium exists there predominantly in oxidation state þIV and to a small part þV, although no signs of V2O5like species have been detected. The VAlF catalysts differ from the VAlO catalysts in their oxygen mobility, which is significantly lower for VAlF, as found with the help of 18O exchange experiments [29]. The high oxygen mobility in VAlO materials is responsible for both a high catalytic activity and predominant total oxidation to CO and CO2, i.e. low selectivity to propene. In the VAlF material there is no mobile lattice oxygen for the total oxidation available; the oxidation obviously occurs via another mechanism at the acidic surface of the catalyst, resulting predominantly in propene formation [29]. As a consequence of these investigations it seems evident that Lewis acidic metal fluorides might be interesting supports for selective oxidation catalysis because of two important advantages that they provide in

86

Functionalized Inorganic Fluorides (a)

XPropane YPropene YCOx

Conversion (X) & yield (Y) (%)

60 50 40 30 20 10 0 0

5

10

15

20

25

30

35

40

V-content (mol-%) (b) 40

XPropan

Conversion (X) & Yield (Y) (%)

35

YPropen YCO

x

30 25 20 15 10 5 0 4

6

8

10

12 14 16 18 V-Content (mol-%)

20

22

24

Figure 3.11 Propane-ODH results with (a) VOx /aluminium oxide (VAlO) and (b) VOx / aluminium fluoride catalysts (VAlF) (from Al:HF ¼ 1:3) with different V contents (773 K; mcat ¼ 100 mg: N2:O2:C3H8 ¼ 20:50:30) (X) conversion; (Y) yield. (Reprinted with permission from [29] Copyright (2008) Elsevier Ltd.)

comparison to all of the metal oxide systems. These are, i) suppression or at the best prevention of any Brønsted acid sites and thus suppression of highly reactive carbocationic intermediates of the organic reactant molecule and ii) prevention of any bulk oxygen mobility in the solid catalyst, which alters the reaction path to a totally different mechanism. The usual Mars and van Krevelen mechanism does not hold for such metal fluoridebased systems. Since similar systems created with the weak Lewis acidic MgF2 system

High Surface Area Metal Fluorides as Catalysts A

B

87

δiso = –167 ppm

–17 ppm

VAlF 15

* 400 kHz

–400 kHz 100

0

0

*

VAlF 15

*

–100 –200 δ (ppm)

* –300

–400

Figure 3.12 27Al MAS NMR spectrum of VAlF (15 mol% V),  rot ¼ 25 kHz, central transition enlarged as insert (left), and 19F MAS NMR of VAlF,  rot ¼ 30 kHz; * spinning side bands (right). (Reprinted with permission from [29] Copyright (2008) Elsevier Ltd.)

V2O5

VOF3

–615ppm

a

–744, –771, –783ppm

c 0

400kHz

–400kHz 1 MHz

0

VAlF 20 350

VAlO 20 350 –582ppm

–797ppm –519ppm

d

b 1 MHz

–1 MHz

0

–1 MHz

1 MHz

0

–1 MHz

Figure 3.13 51V MAS NMR spectra of (a) V2O5 ( rot ¼ 14 kHz), d ¼  615 ppm; (b) VAlO (20 mol% V; calcined at 623 K) ( rot ¼ 13 kHz), d ¼  582 ppm; (c) VOF3 ( rot ¼ 15 kHz), d1 ¼  744 ppm, d2 ¼  771 ppm, d3 ¼  7835 ppm; (d) VAlF (20 mol% V; calcined at 623 K) ( rot ¼ 13 kHz), d1 ¼  519 ppm, d2 ¼  797 ppm. The sharp signal in (b) and (d) is of 27Al because of the very wide sweep width; an overlap of its side bands with those of V can be seen. (Reprinted with permission from [30] Copyright (2007) Elsevier Ltd.)

88

Functionalized Inorganic Fluorides

VAlF 05 VAlF 25 200

250

300

350

400

450

B0 /mT

Figure 3.14 X-band ESR spectra of VOx doped on aluminium fluoride (VAlF 15) and on aluminium oxyfluoride (VOx /AlFyOx) (Reprinted with permission from [29] Copyright (2008) Elsevier Ltd.)

gave just low conversion in these reactions, it proves that Lewis acid sites of a ‘considerable’ strength are needed for these selective oxidation reactions. Unfortunately, quantification cannot be given based on the experimental materials available. Separate oxidation experiments under liquid-phase conditions confirmed the importance of strong Lewis-acid sites on a solid metal fluoride for its performance as oxidation catalyst. HS-AlF3 and HS-Al0.9M0.1F3þd with M ¼ FeIII, MnIII, VIII, and NbIII were prepared in similar way, via sol-gel fluorination, drying and postfluorination with CHClF2 at temperatures up to 503 K, and tested as catalyst for the liquid-phase oxidation of ethylbenzene with tertbutylhydroperoxide (TBHP) [31]. Although HS-AlF3 had the least capacity for O2 chemisorption among the catalysts, it gave a good conversion rate of 42% within 6 h at 333 K; it was only inferior to the Fe and V doped catalysts but was superior to the Mn and Nb doped catalysts. The performance correlates well with the concentration (and strength) of Lewis acid sites, which were determined by the NH3-TPD and IR spectra of adsorbed pyridine. These experiments give evidence that very strong Lewis acid sites on a solid material are able to activate components of oxidation reactions.

3.6

Metal Fluoride Supported Noble Metal Catalysts

There are numerous publications in the literature about noble metal catalysed reactions – too many for us even to try to refer to them here. Noble metals have also been frequently

High Surface Area Metal Fluorides as Catalysts

89

employed for hydrodechlorination reactions of either pure chlorocarbons or chlorofluorocarbons. The synthesis of metal fluoride-supported noble metal catalysts was investigated because of the following intentions: (i) nanoscopic novel metals have attracted considerable attention from chemists, especially the catalytic community, but also among physicists. Many systems have been reported consisting of nanoscopic novel metals supported on metal oxides, very often with superior properties when compared to the classically prepared novel metal catalysts. Hence, it was expected that the fluorolytic sol-gel synthesis of metal fluorides can be readily modified to introduce finely dispersed noble metals like Pt or Pd into the high surface-area metal fluorides, instead loading the noble metal just afterwards onto the metal fluoride as has been almost done so far. It was anticipated that the noble metals might be significantly better dispersed (smaller) when their precursors are already present during the fluorolysis of the metal alkoxides so becoming in situ incorporated into the metal fluoride gel network. (ii) For several halocarbon reactions – e.g. hydrodehalogenation reactions – noble metals are the best catalysts. Such reactions involve both activation of the H-H-bond (at the noble metal site) and activation of the respective C-X-bond (presumably at the Lewis acid site). Thus, metal fluorides should be perfect supports. By using metal fluorides as support, the high thermal and chemical resistance of metal fluorides against HCl and HF can be advantageously employed in reactions where such corrosive conditions are involved (iii) Last but not least, it was intended to apply the obtained nano-nobel metal/nano-metal fluoride catalysts for some other reactions that could benefit from the high Lewis acidity of the fluoride. The noble metals have to be added as, e. g., acetylacetonates in the course of the sol-gel synthesis and after the postfluorination step necessary to bring about the full Lewis acidity of the metal fluoride, a final reduction step follows to convert the noble metal to oxidation state zero. Pd/HS-AlF3, Pd/HS-MgF2 and several other Pd/metal fluoride catalysts as well as Pt/HS-AlF3 and Pt/HS-MgF2, mostly with noble metal content around 5% (w/w), have been prepared, characterized and used as catalysts for hydrodehalogenation reactions and tested for Suzuki coupling reactions. As expected, the synthesis results in a high degree of dispersion for the noble metal (see below) whereas the properties of the respective metal fluoride are only marginally affected. From Figure 3.15 it follows that the X-ray patterns of the noble metal-containing catalysts are almost identical with those of similarly prepared metal fluorides. Thus, Pd/HS-AlF3 is as amorphous as HS-AlF3. It is interesting to note that in course of the postfluorination step the PdII signals diminish, obviously due to the reducing effect of the organic moieties liberated at higher temperatures. Such a reduction is also evident from the respective Pd 3d5/2 XPS binding energies [33]. The surface area and acidity of the noble metal catalysts are comparable although somewhat reduced compared to the respective data of the similarly prepared pure metal fluorides. The performance of these new catalysts and their advantages and disadvantages will be described below based on three different types of reactions, these being i) hydrodechlorination

90

Functionalized Inorganic Fluorides ∗-MgF2

Ο - Pd(acac)2

0

Θ-Pd

II

I

∗

∗ ∗ ∗ ∗

∗

Θ

∗

d

d Ο Ο

Ο Ο Ο

c

Ο Ο

c

Ο ΟΟ

a 40

∆-CaF2

III

50

60

70

a 10

20

30

IV

∆

∆

30

40

50

60

70

-KMgF3

∆

20

∆

10

∆

∆

∆

Intensity/a. u.

b b

∆ ∆

Ο

d

∆

Θ

d Ο Ο Ο

c ΟΟ

ΟΟ

c

Ο

Ο

b b a

10

20

30

40

50

60

70

a 10

2θ/degree

20

30

40

50

60

70

Figure 3.15 X-ray diffraction patterns of (I) AlF3, (II) MgF2, (III) CaF2 and (IV) KMgF3 based catalysts. (a) Pure MFx, (b) PdII/MFx-y(OR)y-precursor, (c) PdII/MFx after post-fluorination and (d) Pd0/MFx after reduction. (Reproduced from [32] by permission of Elsevier.)

of monochlorodifluoromethane, CHClF2, ii) hydrodechlorination of dichloroacetic acid, and iii) Suzuki coupling.

3.6.1

Hydrodechlorination of Monochlorodifluoromethane

After CFCs were phased out, hydrodechlorination of CFCs, especially CFC-12 (CCl2F2) and CFC-114 (CF3-CCl2F), attracted considerable attention from researchers because this was expected to yield as products HFC-32 (CH2F2) and HFC-134 (CF3CH2F), respectively, which are the most important CFC-alternatives. Early work on the hydrodechlorination of CCl2F2 reported that Pt and Pd, supported on alumina and charcoal, respectively, were very reactive catalysts [34–37]. For the hydrodechlorination of CF3-CCl2F, Pd was also found to be a more selective catalyst than Pt, although both systems gave very high yields (cf. [38–40] and references therein). It was found that especially catalyst with -AlF3 as support exhibited best catalytic properties (high conversion and high selectivity), being even better than the best reported Pd catalyst supported on char coal. Hence, we tested the Pd and Pt catalysts described above, supported on highly Lewis acidic nano metal fluorides for the hydrodechlorination of CHClF2 as given in Equation (3.17) aiming the reaction path into CH2F2.

High Surface Area Metal Fluorides as Catalysts CHClF2

+H2 –HCl

CH2F2

+H2

CH3F

–HF

+H2

CH4

91

(3.17)

–HF

Several Pd catalysts supported by AlF3, MgF2, CaF2 or KMgF3 have been prepared, characterized and used for this reaction. The supports were chosen according to their respective Lewis acidity – HS-AlF3 is the second strongest Lewis acid of all, HS-MgF2 is a weak Lewis acid, and CaF2 and, more specifically, KMgF3 are not Lewis acidic. The same order of acidity holds also for the Pd loaded systems. This can be seen from their photoacoustic IR spectra of chemisorbed pyridine of the neat HS-MFxs as well as the respective Pd-loaded ones (Figure 3.16). The surface area of the supported Pd catalysts is about 50% to 75% of the area of the corresponding neat high surface metal fluorides. This might be due to the necessary additional synthesis step, i.e. the reduction of PdII to Pd0 in H2 at 473 K.

A

BPy (1545) LPy + BPy (1490) LPy (1454)

a

Intensity/a.u.

b c d

B a b c d

1575

1550

1525 1500 1475 1450 Wavenumber/cm–1

1425

1400

Figure 3.16 FTIR photoacoustic spectra of pyridine chemisorbed at (A) MFx (a ¼ AlF3, b ¼ MgF2, c ¼ CaF2, d ¼ KMgF3) and (B) Pd0/MFx (a ¼ Pd0/AlF3, b ¼ Pd0/MgF2, c ¼ Pd0/CaF2, d ¼ Pd0/KMgF3). LPy ¼ Lewis acid sites, BPy ¼ Brønsted acid sites. (Reprinted with permission from [32] Copyright (2008) Elsevier Ltd.)

Another important characteristic of HS-MFx-supported Pd catalysts is the high Pd distribution in form of very fine particles. It has been shown by transmission electron microscopy (TEM) of freshly prepared catalyst and of a Pd0/CaF2 catalyst used for CHClF2 hydrodehalogenation reaction that the palladium is integrated as Pd-particles of 3 to 8 nm diameter into CaF2 particles of about 40 to 50 nm diameter (Figure 3.17) as originally anticipated as result of employing the fluorolytic sol-gel-synthesis route.

92

Functionalized Inorganic Fluorides

(a)

(b) 0.31 nm

0.38 nm

2 nm

50 nm (c)

(d) 0.39 nm

50 nm

2 nm

Figure 3.17 TEM images of fresh (a, b) and used (c, d) Pd 0/CaF2 catalysts for hydrodehalogenation of CHClF2. (Reprinted with permission from [32] (2008) Elsevier Ltd.)

The two marked lattice distances correspond to two different phases. Within CaF2 the {111} lattice plane distance of d111 ¼ 0.31(4) was determined and for the metallic Pd the distance of the {100} lattice plane of d100 ¼ 0.38(2). Only these two phases could be detected in these and in other fresh samples investigated, giving evidence for the absence of PdF2 or PdCl2, all Pd is in the metallic state as had already derived from the XRD patterns (see above). In case of the used catalyst (Figures 3.17 c and d) the lattice plane distance of 0.39 nm does not rule out the possible formation of PdF2 or PdO as result of the highly corrosive reaction conditions [33]. Hydrodehalogenation, the selective replacement of halogen by hydrogen on saturated halocarbons, depends on the strength of the C-X bond, the catalyst and the reaction conditions. Thus, upon hydrodehalogenation of CHClF2 C-C coupling products have been observed using Ni catalysts [41], whereas only CH3F and CH4 were obtained with silica-supported Pd [42].

High Surface Area Metal Fluorides as Catalysts

93

Using HS-AlF3 as Pd support, the extremely high Lewis acidity of the support should activate the C-X bond, thus improving catalytic activity and, possibly, selectivity. Appropriate experiments resulted, however, in a product composition not in conformance with that of Equation (3.17) [32]. There was no CH2F2 formed but unexpectedly high amounts of CHF3 (Table 3.3, entry 1). The formation of the latter gives evidence for prevailing dismutation reactions according to Equations (3.18). 2 CHClF2 ! CHCl2 F þ CHF3

(3:18a)

2 CHCl2 F ! CHCl3 þCHClF2

(3:18b)

3 CHClF3 ! 2 CHF3 þCHCl3

(3:18c)

Dismutation has to be catalysed by a Lewis acid of at least moderate strength. The very acidic HS-AlF3 support and even the moderately acidic HS-MgF2 support are not suited to boost the hydrodehalogenation. The high dismutation over these metal fluorides compared to that of HS-CaF2 can be seen from Figure 3.18. AlF3 MgF2 CaF2

Dismutation activity (%)

100 80 60 40 20 0 373

423

473

523

573

623

673

Temperature/K

Figure 3.18 Conversion of CHClF2 over neat HS-MFx. Reaction conditions: amount of catalyst ¼ 0.4 ml, CT ¼ 1 s, GHSV ¼ 3600 h1. (Reprinted with permission from [32] (2008) Elsevier Ltd.)

Consequently, with less Lewis acidic supports, such as CaF2 and KMgF3, hydrodehalogenation prevails over dismutation, as can be seen from the results in Table 3.4. Although with Pd0/CaF2 and Pd0/KMgF3 CHClF2 conversion is very low, the CH2F2 selectivity is significantly higher (around 70%) for both as compared to the 1.7% with the ‘state-of-the-art’ catalyst Pd0/C and its selectivity to CH2F2 of just ca. 5%. The slightly higher conversion (13.6 %) and especially the large amount of other byproducts are probably due to the presence of some metal salts [32].

94

Functionalized Inorganic Fluorides

In conclusion, these investigations clearly prove the potential of the fluorolytic sol-gel synthesis to result in nanoscopic, finely dispersed noble metal catalysts under in situ conditions that are the basis for catalytically highly active metal species. However, the high dismutation reactivity of these catalysts causes undesired side reactions at room temperature, which cannot at all be suppressed. On the other hand, the hydrogenation reactivity significantly reduces when neutral metal fluorides HS-CaF2 are used although particle diameter of palladium and/or platinum lies in the low nanometre scale. This clearly indicates that activation of the C-X-bond in heterogeneously catalysed hydrodehalogenation reactions is obviously the decisive reaction step.

Table 3.3 Hydrodehalogenation of CHClF2 with hydrogen over various metal fluorides supported palladium catalystsa Sr.No

Catalyst

Temp (K)

CHClF2 Conv. (%)

Product selectivity (%) CH2F2

1. 2. 3. 4. 5. 6. 7. a b

Pd0/AlF3 Pd0/MgF2 Pd0/CaF2 Pd0/KMgF3 Pt0/AlF3 Pt0/MgF2 Pd0/C

523 623 623 623 563 563 593

85.9 6.0 2.3 2.9 95.9 0.6 13.6

0 5.2 72.7 68.6 0 0 4.6

CH4

CHF3

CH3F

Otherb

32.1 26.8 27.3 31.4 32.9 48.7 55.5

31.5 39.6 0 0 44.7 51.3 33.1

28.2 6.3 0 0 2.2 0 1.8

8.2 22.2 0 0 20.1 0 5.0

Reaction conditions: amount of catalyst ¼ 0.25mL, H2/CHClF2 mole ratio ¼ 7, GHSV ¼ 3600h-1, CT ¼ 1sec. CH3Cl, CHF2-CHF2, C2H4, etc. (Reprinted with permission from [32] (2008) Elsevier Ltd.)

For hydrodehalogenation reactions in halocarbons with just one kind of halogen in it, e.g. pure fluorocarbons or chlorocarbons, respectively, these new highly Lewis acidic noble metal/MFx-catalysts should therefore be excellent candidates because dismutation as an undesired side reaction can be disregarded. Thus, starting with pure CHCl3 complete hydrodechlorination was observed over Pd0/AlF3 at 423 K whereas the Pd0/Ccatalyst gave only around 70% conversion under the same conditions, with remarkable amounts of incompletely converted mono and dichloromethanes.

3.6.2

Hydrodechlorination of Dichloroacetic Acid (DCA)

Pd/HS-AlF3 and Pd/HS-MgF2 as well the respective Pt-containing catalysts have also been employed for hydrodechlorination of dichloroacetic acid (DCA) to monochloroacetic acid (MCA) according to Equation (3.16), whereby some acetic acid (AA) is also formed. HCl2 CCOOH þ H2 ! H2 ClCCOOH þ HCl

(3:16)

This reaction is important because, upon chloroacetic acid synthesis via chlorination of acetic acid, some dichloroacetic acid is formed as byproduct. Results shown in

High Surface Area Metal Fluorides as Catalysts

95

Table 3.4 give evidence for a good performance of the Pd catalysts with about 50% yield of the aimed MCA; however, about 10% of the converted DCA is reduced to acetic acid (AA).

Table 3.4 Results of catalytic liquid phase hydrodechlorination of dichloroacetic acid* (unpublished results Diploma thesis Pratap Patil, Humboldt-University 2005) Catalyst 5% Pd/HS-AlF3 5% Pt/HS-AlF3 5% Pd/MgF2 5% Pt/MgF2

Reaction Time (h)

DCA Xa (%)

MCA Sb (%)

AA S (%)

MCA Yc (%)

3.5 8 6 8

55 11 55.5 20.6

90.7 99.3 91.6 98.7

9.3 0.7 8.4 1.3

49.9 10.9 50.8 20.3

* Reaction conditions: DCA ¼ 15 mmol (1.934 g), cat ¼ 5 wt% of DCA, Temp ¼ 170 C, H2 flow rate ¼ 20 ml/min. conversion (X). selectivity (S). c yield (Y). a

b

With Pt-based catalysts the selectivity is much better than with Pd-based catalysts, however, the conversion is so low that the MCA yield is much lower than with the Pd catalysts.

3.6.3

Suzuki Coupling

Catalysts employing nano-sized Pd as catalytic active agent are widely used for Suzuki coupling reactions [43]. To make use of the advantages of heterogeneous catalysts the Pd particles have been supported on many different supports [44–48] whereby both small Pd particle size and the properties of the support were found to affect the catalytic activity. To test the suitability of metal fluoride-supported Pd0 catalysts for Suzuki coupling reactions, high surface area AlF3, MgF2 and CaF2 were tested because of their graded Lewis acidity, and additionally different syntheses for Pd0/CaF2 have been tested for cross coupling of 4-bromoanisole (4-BA) with phenylboric acid (PBA) to give 4-methoxybiphenyl (4-MBP) in liquid phase reaction [33]. The catalysts employed for and the results of the Suzuki-coupling experiments are summarized in Table 3.5. The catalytic results in Table 3.5 show the importance of the chemical nature of the support with CaF2 (3a, 3b) being superior to AlF3 (1) and MgF2 (2) even though the latter has the by far highest surface area of the metal fluoride catalysts. Further on, the advantage of the nonaqueous sol-gel fluorination synthesis over aqueous ones becomes apparent by comparing results obtained with 3a, 3b with that obtained with 4. However, all Pd0/CaF2 sol-gel syntheses (3a, 3b, 4, 6) are superior to the ‘classical’ impregnation synthesis (5) as follows from the data in Table 3.5 and in Figure 3.19, most probably because of the very small and homogeneously dispersed Pd particles obtained via sol-gel methods [33].

96

Functionalized Inorganic Fluorides

Table 3.5 Suzuki coupling of 4-bromoanisole (4-BA, I) with phenylboric acid (PBA, II) and K3PO4 (III) over Pd 0/MFx catalysts to 4-methoxybiphenyl (4-MBP) a. (Based on data taken from [33] with permission of the Royal Society of Chemistry.) Catalyst Code

Description

Preparation

1 2 3a 3b 4 5 6 7

Pd0/AlF3-HF Pd0/MgF2-HF Pd0/CaF2-HF Pd0/CaF2-R22 Pd0/CaF2-aq Pd0/CaF2-imp Pd0/Ca(OH)0.5F1.5 Pd0/C

{M(OR)x þ Pd(acac)2 þ aHF; þ HF; þ H2}c {M(OR)x þ Pd(acac)2 þ aHF; þ HF; þ H2} {M(OR)x þ Pd(acac)2 þ aHF; þ HF; þ H2} {M(OR)x þ Pd(acac)2 þ aHF; þ R22; þ H2}d {M(OR)x þ Pd(acac)2 þ HFaq; þ HF; þ H2}e {M(OR)x þ aHF; þ HF; þ Pd(acac)2; þ H2}f {M(OR)x þ Pd(acac)2 þ 1.5aHF; þ H2}g {from Aldrich, Degussa type E101 NO/W)h

SBETb/m2g-1

4-MBP-yield/%

115 225 118 59 37 62 90 –

9 11 88.1 81.9 48.3 8.1 30.6 85.9

a

Reaction conditions: I: II: III ¼ 1: 1.5: 3 (molar ratio), T ¼ 381 K, 35 mg catalyst, solvent: 12 ml of a 5:1 mixture of DMA and H2O. b Specific surface area by BETN2. c Fluorolysis of metal alkoxide and PdIIacetylacetonate in alcohol with stoichiometric amounts of aHF; post-fluorination of the dried product with gaseous HF; reduction of Pd(acac)2 with H2. Pd content of all samples adjusted to 5 wt%. d like c except: postfluorination with gaseous CHClF2. e like c except: fluorolysis with 40 % aqueous HF. f Preparation of high surface area CaF2 (S ¼ 99 m2 g-1) followed by impregnation with Pd(acac)2 in CHCl3 and H2 reduction of the dried sample. g Fluorolysis of methanolic solution of Ca(OMe)2 and Pd(acac)2 with 1.5 equivalents of aHF, followed by drying and calcinations in air at 350 C, followed by H2 reduction at 200 C. h Commercial catalyst for comparison.

Yield of 4-methoybenzophenone (%)

100 80 60

Pd/CaF2-HF(3a) Pd/CaF2-R22(3b) Pd/CaF2-Aq(4) Pd/CaF2-imp(5)

40 20 0 0

2

4 6 Reaction time/h

8

10

Figure 3.19 Catalytic activity of Pd 0/CaF2 samples for Suzuki coupling of 4-bromoanisole with phenylboric acid. (Reprinted with permission from [33] Copyright (2008) Royal Society of Chemistry.)

High Surface Area Metal Fluorides as Catalysts

97

Thus, even for the Suzuki coupling reaction, noble metals supported on nanoscopic metal fluorides obtained via the fluorolytic sol-gel-route are very active and selective although there is just a small advantage over the ‘state-of-the-art’ Pd0/Ccatalyst. It might be expected, however, that the application of this new class of noble metal catalyst would be most advantageous for reactions where at least one reaction step requires activation at a Lewis-acid site. Hence, many such reactions have still to be evaluated. For some of them these catalysts are expected to be superior to ‘classical’ oxide-based catalysts. In conclusion, in several catalysed reactions the fluorolytic sol-gel synthesis-derived nanoscopic metal fluorides have already shown strong potential. Although it is probably the case that, for many catalytic reactions, metal oxides might remain the preferred options for many reasons, there are several very promising fields for which nanoscopic metal fluoride-based catalysts are advantageous. For many halogen exchange reactions, in particular, these new catalysts exhibit unique properties.

References [1] K.O. Christe, D. A. Dixon, D. McLemore, W. W. Wilson, J. A. Sheehy and D. A. Boatz, On a quantitative scale for Lewis acidity and recent progress in polynitrogen chemistry, J. Fluorine Chem., 101, 151–153 (2000). K. O. Christe, Quantitative Lewis Acidity Scale, a Progress Report, 19th Winter Fluorine Conference, St. Petersburg (USA) 2009, Abstract 59. [2] J. Krishna Murthy, U. Groß, St. Ru¨diger, E. Kemnitz and J.M. Winfield, Sol-gel-fluorination synthesis of amorphous magnesium fluoride, J. Solid State Chem., 179, 739–746 (2006). [3] St. Ru¨diger and E. Kemnitz, The fluorolytic sol-gel route to metal fluorides – a versatile process opening a variety of application fields, Dalton Trans., 1105–1252 (2008). [4] St. Ru¨diger, G. Eltanany, U. Groß and E. Kemnitz, Real sol-gel synthesis of catalytically active aluminium fluoride, J. Sol-Gel Sci. Techn., 41, 299–311 (2007). [5] J. Krishna Murthy, U. Groß, St. Ru¨diger, V. Venkat Rao, V. Vijaya Kumar, A. Wander, C. L. Bailey, N. M. Harrison and E. Kemnitz, Aluminium chloride as a solid is not a strong Lewis acid, J. Phys. Chem. B, 110, 8314–8319 (2006). [6] H. Bozorg Zadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, J. M. Winfield, J. Fluorine Chem., 107 (2001) 45/52. [7] St. Ru¨diger, U. Groß, M. Feist, H. A. Prescott, S. Chandra Shekar, S. I. Troyanov and E. Kemnitz, Non-aqueous synthesis of high surface area aluminium fluoride – a mechanistic investigation, J. Mater. Chem., 15, 588–597 (2005). [8] H. Bozorg Zadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin, G. D. Tate, J. M. Winfield, J. Fluorine Chem. 112 (2001) 225/ 32. [9] M. Nickho-Amiry, J.M. Winfield, Investigation of fluorinated surfaces by means of radiolabelled probe molecules, J. Fluorine Chem. 128 (2007) 344–352 [10] G. Eltanany, St. Ru¨diger and E. Kemnitz, Supported high surface AlF3: a very strong Lewis acid for catalytic applications, J. Mater. Chem., 18, 2268–2275 (2008). [11] K. Tanabe, T. Sumiyoshi, K. Shibata, T. Kiyoura and J. Kitagawa, A new hypothesis regarding the surface acidity of binary metal oxides, Bull. Chem. Soc. Japan 47 (1974) 1064–1066. [12] E. Kemnitz, Y. Zhu and B. Adamczyk, Enhanced Lewis Acidity by Aliovalent cation doping in metal fluorides, J. Fluorine Chem. 114 (2002) 163–170. ¨ nveren and E. Kemnitz, Mixed metal fluorides as [13] J. Krishna Murthy, U. Groß, St. Ru¨diger, E. U doped Lewis acidic catalysts systems: a comparative study involving novel high surface area metal fluorides, J. Fluorine Chem., 125, 937–949 (2004).

98

Functionalized Inorganic Fluorides

[14] J. Krishna Murthy, U. Groß, St. Ru¨diger and E. Kemnitz, FeF3/MgF2: novel Lewis acidic catalyst systems, Appl. Catal., A, 278, 133–138 (2004). ¨ nveren, W. Unger and E. Kemnitz, Synthesis and [15] J. Krishna Murthy, U. Groß, St. Ru¨diger, E. U characterization of chromium(III)-doped magnesium fluoride catalysts, Appl. Catal., A, 282, 85–91 (2005). [16] H. A. Prescott, Z.-J. Li, E. Kemnitz, J. Deutsch and H. Lieske, New magnesium oxide fluorides with hydroxy groups as catalysts for Michael additions, J. Mater. Chem., 15, 4616–4628 (2005). [17] K. A. Parker and T. L. Mindt, Electrocyclic ring closure of the enols of vinyl quinones. A 2H-chromene synthesis, Org. Lett., 3, 3875–3878 (2001). [18] St. Wuttke, S. M. Coman, G. Scholz, H. Kirmse, A. Vimont, M. Daturi, S. L. M. Schroeder and E. Kemnitz, Novel sol-gel Synthesis of acidic MgF2-x(OH)x materials, Chem. Eur. J. 14, 11488/ 11499 (2008). [19] A. Hasegawa, K. Ishihara and H. Yamamoto, Trimethylsilyl pentafluorophenylbis(trifluoromethanesulfonyl)methide as a super Lewis acid catalyst for the condensation of trimethylhydroquinone with isophytol, Angew. Chem. Int. Ed., 42, 5731–5733 (2003). [20] S.M. Coman, S. Wuttke, A. Vimont, M. Daturi and E. Kemnitz, Catalytic performance of nanoscopic - AlF3 based catalysts in the synthesis of (all-rac)--tocopherol, Adv. Synth. Catal, 350, 2517–2524 (2008). [21] S.M. Coman, P. Patil, St. Wuttke and E. Kemnitz, Cyclisation of citronellal over heterogenous inorganic fluoride – highly chemo- and diasterioselective catalysts for (–)-isopulegol, Chem. Commun. 460–462 (2009). [22] E. A. Mamedov and V. Corte´s Corbera´n, Oxidative dehydrogenation of lower alkanes on vanadium oxide-based catalysts. The present state of the art and outlooks, Appl. Catal., A, 127, 1–40 (1995). [23] M.M. Bettahar, G. Costentin, L. Savary and J. C. Lavalley, On the partial oxidation of propane and propylene on mixed metal oxide catalysts, Appl. Catal., A, 145, 1–48 (1996). [24] Y. Habuta, N. Narishige, K. Okumura, N. Katada and M. Niwa, Catalytic activity and solid acidity of vanadium oxide thin layer loaded on TiO2, ZrO2, and SnO2, Catal. Today, 78, 131–138 (2003). [25] I. E. Wachs, J.-M. Jehng, G. Deo, B. M. Weckhuysen, V. V. Guliants and J. B. Benziger, In situ Raman spectroscopy studies of bulk and surface metal oxide phases during oxidation reactions, Catal. Today, 32, 47–55 (1996). [26] A. Pantazidis, A. Auroux, J.-M. Herrmann and C. Mirodatos, Role of acid–base, redox and structural properties of VMgO catalysts in the oxidative dehydrogenation of propane, Catal. Today, 32, 81–88 (1996). [27] K. Scheurell, E. Hoppe, K.-W. Brzezinka and E. Kemnitz, Bulk and surface properties of highly dispersed VOx/ZrO2, VOx/SiO2 and VOx/TiO2/SiO2 systems and their relevance for propane oxidation, J. Mater. Chem., 14, 2560–2568 (2004). [28] K. Scheurell and E. Kemnitz, Amorphous aluminium fluoride as new matrix for vanadiumcontaining catalysts, J. Mater. Chem., 15, 4845–4853 (2005). [29] K. Scheurell, G. Scholz, A. Pawlik and E. Kemnitz, VOx doped Al2O3 and AlF3 – a comparison of bulk, surface, and catalytic properties, Solid State Sci., 10, 873–883 (2008). [30] K. Scheurell, G. Scholz and E. Kemnitz, Structural study of VOx doped aluminium fluoride and aluminium oxide catalysts, J. Solid State Chem., 180, 749–758 (2007). [31] I. K. Murwani, K. Scheurell, E. Kemnitz, Liquid phase oxidation of ethylbenzene on pure and metal doped HS-AlF3, Cat. Com. 10, 227–231 (2008). [32] P. T. Patil, A. Dimitrov, H. Kirmse, W. Neumann and E. Kemnitz, Non-aqueous sol-gel synthesis, characterization and catalytic properties of metal fluoride supported palladium nanoparticles, Appl. Catal., B, 78, 80–91 (2007). [33] B. Coq, J. M. Cognion, F. Figueras and D. Tornigant, Conversion under hydrogen of dichlorodifluoromethane over supported palladium catalysts, J. Cat. 141, 21–33 (1993). [34] Z. Karpinski, K. Early and J. L. d’Itri, Catalytic Hydrodechlorination of 1,1-Dichlorotetrafluoroethane by Pd/Al2O3, J. Cat. 164, 378–386 (1996). [35] E. J. A. X. van de Sandt, A. Wiersma, M. Makkee, H. van Bekkum and J. A. Moulin, Palladium black as model catalyst in the hydrogenolysis of CCl2F2 (CFC-12) into CH2F2 (HFC-32), Appl. Cat. A 155, 59–73 (1997).

High Surface Area Metal Fluorides as Catalysts

99

[36] E. J.A X. van de Sandt, A. Wiersma, M. Makkee, H. van Bekkum and J.A. Moulin, Process for the selective hydrogenolysis of CCl2F2 (CFC-12) into CH2F2 (HFC-32), Cat. Today 27, 257–264 (1996). [37] Z. Karpinski, J.L. d’Itri, Hydrodechlorination of 1,1-dichlorotetrafluoroethane on supported palladium catalysts, a static-circular reactor study, Catal. Lett. 77, 135–140 (2001). [38] H. Berndt, H. Bozorg Zadeh, E. Kemnitz, M. Nickkho Amiry, M. Pohl, T. Skapin and J. M. Winfield, The properties of platinum or palladium supported on b-aluminium trifluride or magnesium dfluoride: catalysts for the hydrodechlorination of 1,1-dichlorotetrafluoroethane, Mat. Chem. 12, 3499–3507 (2002). [39] I. Kris-Murwani, E. Kemnitz, T. Skapin, M. Nickkho-Amiry and J. M. Winfield, Mechanistic investigation of the hydrodechlorination of 1,1,1,2-tetrafluorodichloroethane on metal fluoride-supported Pt and Pd, Ctal. Today 88, 153–168 (2004). [40] P. T. Patil, A. Dimitrov, J. Radnik and E. Kemnitz, Sol-gel synthesis of metal fluoride supported Pd catalysts for Suzuki coupling, J. Mater. Chem., 18, 1632–1635 (2008). [41] A. Morato, C. Alonso, F. Medina, P. Salagre, J. E. Sueiras, R. Terrado and A. Giralt, Conversion under hydrogen of dichlorodifluoromethane and chlorodifluoromethane over nickel catalysts, Appl. Catal., B, 23, 175–185 (1999). [42] R. Hina, I. Arafa and A. Masadeh, Hydrogenation of CHClF2 (CFC-22) over Pt-supported on silicabased polymethylsiloxane composite matrices, React. Kinet. Catal. Lett., 87, 191–198 (2005). [43] N. Miyaura and A. Suzuki, Palladium-catalyzed cross-coupling reactions of organoboron compounds, Chem. Rev., 95, 2457–2483 (1995). [44] D. D. Das and A. Sayari, Applications of pore-expanded mesoporous silica. 6. Novel synthesis of monodispersed supported palladium nanoparticles and their catalytic activity for Suzuki reaction, J. Catal., 246, 60–65 (2007). [45] F.-X. Felpin, T. Ayad and S. Mitra, Pd/C: An old catalyst for new applications – its use for the Suzuki–Miyaura reaction, Eur. J. Org. Chem., 2679–2690 (2006). [46] K. Shimizu, R. Maruyama, S. Komai, T. Kodama and Y. Kitayama, Pd-sepiolite catalyst for Suzuki coupling reaction in water: structural and catalytic investigations, J. Catal., 227, 202–209 (2004). [47] B. M. Choudary, S. Madhi, N. S. Chowdari, M. L. Kantam and B. Sreedhar, Layered double hydroxide supported nanopalladium catalyst for Heck-, Suzuki-, Sonogashira-, and Stille-type coupling reactions of chloroarenes, J. Am. Chem. Soc., 124, 14127–14136 (2002). [48] M.L. Kantam, K. B. S. Kumar, P. Srinivas and B. Sreedhar, Fluoroapatite-supported palladium catalyst for Suzuki and Heck coupling reactions of haloarenes, Adv. Synth. Catal., 349, 1141–1149 (2007).

4 Investigation of Surface Acidity using a Range of Probe Molecules Alexandre Vimont1, Marco Daturi1 and John M. Winfield2 1

Laboratoire Catalyse et Spectrochimie, ENSICAEN, Universite´ de Caen, CNRS, 6 Bd Mare´chal Juin, F-14050 Caen, France 2 Department of Chemistry, University of Glasgow, Glasgow G12 8QQ, UK

4.1

Introduction

Acidity and acid-base reactions are related and important features of the chemistry of molecular fluorides. In addition to the archetypal Brønsted acid anhydrous hydrogen fluoride (I), molecules in which an hydroxyl group is bonded to a high oxidation state centre (II), for example TeVI, also behave as strong Brønsted acids. The coordination numbers found most commonly in high oxidation state binary fluorides are four and six, hence important molecular Lewis acids are (III), (IV) and (V), where X can be B, Si or Sb for example, which in many situations behave as coordinatively unsaturated species. The Brønsted and Lewis acidities of these high oxidation state centres are enhanced by the high electronegativity of F; as a result, these species are used very widely in solution as reagents or molecular acidic catalysts.

Functionalized Inorganic Fluorides: Synthesis, Characterization & Properties of Nanostructured Solids  2010 John Wiley & Sons, Ltd

Edited by Alain Tressaud

102

Functionalized Inorganic Fluorides H O F

F

X

H–F F

F

X

F F

F F

(I)

(II)

X

4.1.1

F F

(IV)

F

F

F

F

(III)

F

X

F

F

F

F

F

X

F F

(V)

Setting the Scene: Metal Fluorides versus Metal Oxides

The situation is very different in the solid state. The most widely used class of solid acid reagents or catalysts is not fluorides but metal oxides, such as silicas, the transitional aluminas, aluminosilicates and high oxidation state transition metal oxides, which include titania and zirconia. One of the important properties of these materials is their relatively large specific surface areas; in contrast, the analogous metal fluorides, as conventionally prepared, have surface areas that are determined by purely geometric factors and are therefore small. Traditional solid-state synthesis methods lead to the production of such small surface area samples, whose powders are constituted by large crystallites. Conversely, soft chemistry methods, which were not applied to metal fluoride synthesis until relatively recently, allow nanosized particles to be obtained having unconstructed surfaces. Additionally, metal fluorides are extremely sensitive to traces of water. Probably due to a combination of specific surface area and susceptibility to hydrolysis, heterogeneous acid catalysts for use in large-scale fluorination processes are usually based on either chromia or -alumina, which become fluorinated in situ. In laboratory situations, metastable aluminium trifluoride phases having structures that are more open than the close-packed -AlF3 show some promise as acid catalysts.1 Other types of functionalized oxides, which have acidic properties and therefore potential catalytic value, include sulfated oxides2 (for a recent laboratory application in organofluorine chemistry see reference 3) and fluorinated clays.4 In recent work, much of which is reviewed elsewhere in this volume, surface areas of metal fluorides, for example as measured by the BET method, can be increased markedly if the synthetic method leads to aggregates of very small (sometimes called nanoscopic) particles. Such solids will exhibit properties that are very different from the conventionally prepared analogues, for example a molecular dynamics simulation of cubic nanoparticles of -AlF3 indicated the presence of structural motifs that are found also in the metastable

Investigation of Surface Acidity

103

phases such as b-AlF3.5 Most importantly in the context of this chapter, is the presence of a large number of four or five coordinate AlIII centres on the surface of the nanoparticles, consistent with the generation of significant Lewis acidity. In contrast, the thermodynamically stable -AlF3, as conventionally prepared, shows little or no Lewis acid character. Recent progress in the development of synthetic routes to metal fluorides, particularly those of aluminium and magnesium, which have high surface areas, has been described in earlier chapters of this volume. The significant catalytic properties of these high surface area (HS) materials have also been emphasized. Here we examine the surface acidities, both Lewis and Brønsted, of newly synthesized fluorides based on AlIII and MgII centres and pose the question ‘What is meant by a strong acid in an heterogeneous context?’

4.1.2

Some Examples of the Application of FTIR Spectroscopy to the Study of Surface Acidity in Metal Oxides

It is convenient to use metal oxides as reference points to compare the surface acidity of metal fluorides as oxides have been studied very widely and much of the methodology described for fluorides below was developed from studies of high surface area oxides. Some examples of studies made using FTIR spectroscopy serve to introduce this aspect of the subject. A method that is common in the study of surface Lewis acidity of aluminas is to use the Lewis base, pyridine (py), monitoring the IR spectrum of the chemically adsorbed species. Observable shifts in some vibrational modes, in particular  8a and  19b, occur on coordination at the surface.6,7 Adsorption of py at activated -alumina, similar structurally to the commonly studied acid, -alumina, was followed by desorption over a range of temperatures, using diffuse reflectance infrared Fourier Transform spectroscopy (DRIFTS) to monitor the surface. Observation of the py 8a mode band intensity at various temperatures was in accord with the presence of three types of Lewis acid surface sites. This conclusion was refined by studying temperature programmed desorption of py under different conditions.8 The desorption process associated with py bound to medium strength sites comprises two components, whose relative intensities are approximately 1:1. These, and the strong Lewis acid sites, can be quantified by gravimetry. In order to propose descriptions at the ‘molecular’ level for the four types of acid site, use is made of the changes observed in the –OH stretching region of -alumina as py is adsorbed and then desorbed. It is possible therefore to correlate the relative strengths of the Lewis acid surface sites, as determined from the thermal behaviour of the py 8a vibrational modes, with the temperature-dependent changes observed in the –OH stretching region. By using the proposals made previously in the literature for the various terminal –OH group environments on aluminas,9 schematic descriptions, Figure 4.1, can be generated, which take into account the interactions between coordinated py and near neighbour surface –OH groups, depicted in Figure 4.2. An example of this methodological approach is depicted in Figure 4.3, which presents spectra of pyridine resulting from adsorption at a -alumina sample, synthesized by

104

Functionalized Inorganic Fluorides H O

H O ∗ AIOct

AIOct

∗ AITet

AIOct

AIOct i) Weak Lewis Acid Site

ii) Medium-Weak Lewis Acid Site

OH

∗ AITet

∗ AITet

AITet O

iii) Medium-Strong Lewis Acid Site

iv) Strong Lewis Acid Site

Figure 4.1 Schematic of the four types of surface Lewis site proposed for Z-alumina. (Reprinted with permission from D. T. Lundie, A. R. McInroy, R. Marshall, J. M. Winfield, P. Jones, C. C. Dudman, S. F. Parker, C. Mitchell and D. Lennon, J. Phys. Chem., B, 109, 11592–11601 Copyright (2005) American Chemical Society.)

H O AIOct

AIOct

H O

N

N

AIOct

AITet

AIOct i) Weak Lewis Acid Site

OH AITet

ii) Medium-Weak Lewis Acid Site

N

N AITet

AITet

O

iii) Medium-Strong Lewis Acid Site

iv) Strong Lewis Acid Site

Figure 4.2 Schematic of the pyridine complexes proposed for the four types of surface Lewis acid site in Z-alumina. (Reprinted with permission from D. T. Lundie, A. R. McInroy, R. Marshall, J. M. Winfield, P. Jones, C. C. Dudman, S. F. Parker, C. Mitchell and D. Lennon, J. Phys. Chem., B, 109, 11592–11601 Copyright (2005) American Chemical Society.)

Investigation of Surface Acidity

105

Rhodia (100 m2 g1). Spectrum a shows the behaviour of the sample as is, before any activation treatment. The large ‘bump’ between 3800 and 2500 cm1 is due to water adsorbed on the surface, while the features in the range 1700–1300 cm1 belong to carbonate impurities. After a thermal treatment under dioxygen, then secondary vacuum (104 Pa), only residual OH groups remain as shown by bands between 3750 and 3600 cm1 (spectrum b). Spectrum c shows the result of pyridine adsorption then evacuation on the activated sample. C–H stretches of the adsorbed molecules give rise to the features at 3200–2900 cm1,

absorbance

= 0.5 a.u.

(a) (b) (c) (d)

4000

3500

3000 2500 2000 wavenumber/cm–1

1500

= 0.2 a.u.

absorbance

(a) (b)

ν 19b ν 8a ν 8b

ν 19a

(c) (d) 1700

1650

1600 1550 1500 wavenumber/cm–1

1450

Figure 4.3 Transmission IR spectra of pyridine adsorbed on g-alumina: top, MID-IR absorption region; bottom, zoom on the ring vibration region. (a) sample before activation; (b) activated sample; (c) pyridine adsorption and evacuation at room temperature; (d) difference between c and b spectra

106

Functionalized Inorganic Fluorides

whereas ring vibrations produce bands in the 1650–1400 cm1 range. For the sake of clarity, the difference between spectra recorded after pyridine adsorption and after activation is given in spectrum d. The bottom part of Figure 4.3 represents a zoom of the image presented above, in the most interesting spectral region, i.e. in the range 1700–1400 cm1. The main vibrational modes of the pyridine ring are apparent. From bands  8a it is possible to get information on the strength of the acid sites (the higher the wavenumber, the stronger the acidity), whereas the integrated intensity of the  19b furnishes a quantitative estimation of the acid sites concentration. Further details are reported in references 10 and 11. When H-bonded species are concerned, the corresponding bands for pyridine are less sensitive and closer to the frequencies observed in the liquid phase. Such a small frequency shift can make assignments ambiguous when both weak Lewis and Brønsted acid sites are expected on the surface. In such a case, the only way to prove the formation of H-bonded species is the concomitant observation of a broad (OH) band, downward shifted by several hundreds of wavenumbers from the frequency of the free OH groups. This observation is not always straightforward because this band can become very broad and overlap with (CH) bands. Moreover, for weak interactions, the  8a band is in a Fermi resonance with the  1 þ  6a combination mode, so introducing an additional difficulty to assess the acid strength of the acid centres. To circumvent these problems it was found that an accurate analysis of the whole spectrum of pyridine, including the (CH) range, may bring further insights to the acidic properties of metal oxides. Using d5-pyridine, in particular, allowed simpler (CD) spectra to be obtained, showing that both frequencies and intensities are sensitive to the adsorption mode and making it possible to distinguish among H-bonded, protonated and coordinated species.10 Particular attention was paid to the case of coordinated species, since the strength of coordination increases (CH/D) frequencies and decreases their intensities. Such variations were investigated by DFT calculations and explained by the polarization of electron density towards nitrogen to the detriment of hydrogen atoms, leading to (i) an increase of C–H bond strength and (ii), a decrease of the dipole moment derivatives for C–H stretches. These trends were correlated with the  8a and  19b frequency shifts (Figure 4.4). It appeared that the most intense (CH) band is more sensitive to small acidity differences than those in the (C¼C) range, allowing Lewis acid strength to be discriminated even for solids having a similar acidic behaviour, such as CaO, MgO and CeO2. In many circumstances, the use of a single probe molecule is not sufficient to obtain a satisfactory view of the solid surface. In such a case the combination of different probe molecules is recommended. For example, the study of surface properties of -Ga2O3, an oxide structurally related to -Al2O3, was undertaken using pyridine, dimethyl pyridine (lutidine, DMP) (for acidity), CD3CN and carbon dioxide (for basicity), and compared with the vibrations of OH groups, which can be considered as a probe molecule intrinsic to the solid itself.11 Among the different properties observed it is noteworthy that, similar to what is known for -Al2O3, the Brønsted acidity of (partially hydroxylated)

-Ga2O3 was found to be very low, albeit not negligible. DMP adsorbed on -Ga2O3 and

-Al2O3 showed, in both cases, a weak IR absorption band at 1649–1652 cm1 that identified the protonated DMPHþ species. However, no traces of the pyridinium ion were found when pyridine was adsorbed on either -Ga2O3 or -Al2O3. Moreover, the surface Lewis acid strength was found to be slightly smaller in -Ga2O3 than in

Investigation of Surface Acidity ν(CD)

ν(CH) 3105

2340

3100

2330

3095

2320

3090 3085

(cm–1)

(cm–1)

107

3080 3075

2310 2300 2290

3070

2280

3065 3060 1437 1442 1447 1452 1457 1462 1467 ν19b (cm–1)

2270 1290 1300 1310 1320 1330 1340 1350 ν19b (cm–1)

Figure 4.4 Frequency of the most intense n(CH/D) band versus the frequency of the n19b mode of h5-pyridine (left) and d5-pyridine (right) coordinated on various metal oxides. (Reprinted with permission from A. Travert, A. Vimont, A. Sahibed-Dine, M. Daturi, J.-C. Lavalley, Appl. Catal. A: General, 307, 98–107 Copyright (2006) Elsevier Ltd.)

-Al2O3, due to the smaller polarizing power of coordinatively unsaturated GaIV ions, which have a smaller charge/radius ratio than Al3þ. In contrast, the concentration of Lewis acid sites (detected by adsorbed pyridine and DMP) was found to be higher in

-Ga2O3, thanks to the higher tetrahedral preference of the Ga3þ ion. Both -Ga2O3 and

-Al2O3 showed a similar surface basicity toward adsorbed carbon dioxide and acetonitrile. However, the surface concentration of basic O2 ions was found to be larger for gallium oxide, as deduced from the larger amount of acetamide species formed upon adsorption of acetonitrile.

4.1.3

A Preview

Having set the scene with some descriptions of surface chemistry of some metal oxides, we shall turn to the ambiguities in surface acidity determinations compared with the now rather precise acidity determinations that are possible for molecular fluorides. A possible benchmark for solid fluoride Lewis acids is aluminium chlorofluoride. Experimental methodology is then described, both the familiar FTIR method, developed directly from oxide studies (cf. § 4.1.2) and the less familiar use of chlorine-36 labelled anhydrous hydrogen chloride, used either directly or generated at a surface by dehydrochlorination of tert-butyl chloride, also labelled with [36Cl]. The use of both techniques for the investigation of various recently synthesized AlIII and MgII fluorides is described separately; finally a multi-probe approach, involving both techniques, is used to investigate the rather complex surface acidity of the hexagonal tungsten bronze aluminium hydroxy fluoride.

108

Functionalized Inorganic Fluorides

4.2

Characterization of Acidity on a Surface: Contrasts with Molecular Fluorides

One of the driving forces behind recent developments in the synthesis and characterization of new acids is the requirement to replace mineral acids by reagents and catalysts that will lead to reduction in waste, in other words to an improved atom economy. Thus, HCl used in the synthesis of diamino diphenyl methane, an intermediate in polyurethane production, can be replaced by a silico-aluminate catalyst in which the acid sites of the zeolite are accessible through the external surface.12 Less conventionally, simple aromatic hydrocarbons can be nitrated using 60–70% nitric acid in the presence of the strong Lewis acids, lanthanide-perfluoroalkane sulfonic acid salts, such as the ytterbium(III) compounds (VI) and (VII).13 Yb(OSO2CF3)3

Yb(OSO2C8F17)3

Yb[C(SO2RF)3]3

Hf[N(SO2C8F17)2]4

(VI)

(VII)

(VIII)

(IX)

The class of molecular Lewis acids, encompassing f and d block metal derivatives of the strong Brønsted acids, HC(SO2RF)3 and HN(SO2RF)2, for example (VIII) and (IX), finds extensive use in laboratory synthesis, particularly Friedel Crafts acylation14 and other acid catalysed reactions of organics.15 Separation techniques are crucial, biphasic fluorous methods often being used where the RF group is C8 or greater. Supporting the Lewis acids on a polymer is another possibility. Although these catalysts have been very successful in small-scale syntheses, the environmental persistence of molecules containing long chain perfluorocarbon groups would be a disadvantage for large scale processes. Cost is another factor that favours the use of oxides and possibly binary fluorides as acids.

4.2.1

Molecular Brønsted and Molecular Lewis Acids

Molecular Brønsted acids have been studied widely in a variety of liquid media following the pioneering work of R. J. Gillespie and his school.16 Fluorine-containing acids are used widely and provide some of the strongest Brønsted acids known. Quantitatively, acidity is measured by determination of the Hammett acidity function, Ho (X). There are several spectroscopically based methods for its determination, making use of series of aromatic hydrocarbon derivative weak bases, B, which are partially protonated in solution.17 B + H+

BH+ Ho = pKBH+ – log[BH+] / [BH] (X)

The area of super acidic media, defined as media with Ho values more negative than that of 100% H2SO4 (–11.9) has been reviewed fairly recently, with emphasis on fluorinecontaining acids.18 Unlike the situation for Brønsted acidity, manifest by complete or partial proton transfer from acid to base, Lewis acidity is a more general concept. It encompasses the BrønstedLowry definition and thus there is no universal scale (in contrast to the Hammett acidity

Investigation of Surface Acidity

109

scale). Many alternative scales have been devised. Acidity and basicity cannot be related to a ‘universal’ e pair donor and relative orders of Lewis acidity will depend on the identity of the reference base used. Ideally, since an acid-base interaction is equivalent to bond energy, strengths of Lewis acids versus a chosen base should be determined by calorimetry, or an equivalent method that yields a thermodynamic quantity. The experiment should be carried out with isolated molecules, usually translated as being in the gas phase or in a nonsolvating solvent (if this latter is not a contradiction in terms). Relatively few systems have been studied in this way, hence the widespread use of computations using DFT methods and of spectroscopic methods leading to relative shifts (IR or NMR are the most popular); these types of approaches to the relative Lewis acidities of the heavier Group 3 halides have been compared.19 Comparisons among molecular Lewis acid fluorides are often made on the basis of the F ion affinities of the isolated molecules, for example using the pF scale.20 The ordering of Lewis acidities is based largely on computational work, the values being referenced to the experimentally determined gas phase F ion affinity of OCF2 (XI). OCF2 (g) + F– (g)

OCF3–(g) (XI)

From these computations and related work,21, 22 it has been established quantitatively that antimony pentafluoride and its oligomers are benchmarks for molecular, strong Lewis acids.

4.2.2

A Possible Benchmark for Solid Metal Fluoride, Lewis Acids: Aluminium Chlorofluoride

The process of making comparisons among solid Lewis or Brønsted acids is more complicated than for the molecular acids illustrated above. For example the use of Hammett indicators to determine Brønsted acidity in sulfated oxides has proved to be controversial;2 the values obtained can be misleading because heterogeneous conditions are very different from those in a nominally ideal solution. Not only are the intrinsic strengths of the different types of surface acid important but surface site densities and morphology (the latter could be in conflict with the steric requirements of the species used as a probe) may be crucial in determining the apparent effectiveness of a solid acid. In the context of fluorine chemistry, replacement of the strong Lewis acid, liquid SbF5, by a strong, solid Lewis acid, which is easily separable and thus reusable, would result in significant waste and energy reductions. The hypothetical, monomeric molecule AlF3 has a high pF value from the DFT computation20 and for this reason, development of new aluminium fluorides as potentially strong Lewis acids has attracted particular attention. Solid aluminium chlorofluoride (ACF)19, 23 was an early candidate, notwithstanding its extremely hygroscopic nature, and in some types of reaction its behaviour resembles that of liquid SbF5. For example, both compounds behave as catalysts for the electrophilic addition of fluoroalkanes to perfluoroolefins, although the former is the catalyst of choice using halopolyfluoroalkanes,24 while the latter is more effective using hydrofluoroalkanes.25

110

Functionalized Inorganic Fluorides

Aluminium chlorofluoride is an amorphous solid of high specific surface area. It is prepared by halogen exchange between aluminium trichloride and a chlorofluorocarbon under conditions where exchange is almost complete; its composition is AlClxF3x, x ¼ approximately 0.05–0.3.26 A structure has been proposed from data analyses involving several physical methods.19,27 Chlorine is an integral part of the environment about octahedrally coordinated AlIII (there is no evidence in this composition range for an AlIII chloride phase); long range order among the various structural motifs is absent; those proposed are (XII): [Al(mF)5/2F] 3[Al(m3Cl)1/3(mF)5/2] n[Al(mF)6/2] : n is variable. (XII) Although there is no specific information available regarding the exact composition of the surface, it is logical to assume that coordinatively unsaturated AlIII sites are present having a variety of F, Cl environments. The compound ACF is the benchmark for the [36Cl]-radiotracer approach to Lewis acidity which is described below. The compound has therefore been described in some detail.

4.3 4.3.1

Experimental Methodology FTIR Spectroscopy

The in situ characterization of a solid surface by FT-IR is a well known and current method of investigation in a number of laboratories. Nevertheless, each team of scientists may use different analysis techniques. In our case we choose the transmission technique, so that samples are pressed (109 Pa) into self-supported discs (2 cm2 area, 7–10 mg cm2). They are placed in a home-made quartz cell equipped with KBr windows (Figure 4.5). A movable quartz sample holder permits adjustment of the pellet in the infrared beam for spectral acquisition and its displacement into a furnace at the top of the cell for thermal treatments. The cell is connected to a vacuum line for evacuation, calcination steps (residual pressure ¼ 103–104 Pa) and for the introduction of probe molecules (in gas or vapour phase) into the infrared cell. Spectra are recorded at room temperature. In adsorption experiments where probe molecules having weak interactions are used (such as CO), the temperature of the pellet is decreased to about 100 K by cooling the sample holder with liquid N2 after quenching the sample from the thermal treatment temperature. The addition of accurately measured increments of probe molecules in the cell (a typical increment corresponds to 100 mmol of probe per g of material) is possible via a calibrated volume (1.50 cm3) connected to a pressure gauge for the control of the probe pressure (1–104 Pa range). The probe pressure inside the IR cell is monitored by another pressure gauge (1–103 Pa range). Transmission IR spectra are recorded in the 5600–500 cm1 range, at 4 cm1 resolution, on a Nicolet Nexus spectrometer equipped with an extended KBr beam splitting device and a mercury cadmium telluride (MCT) cryodetector. Prior to the adsorption of probe molecules, the samples are activated at different temperatures (ranging from 373 to 873 K, depending on the material and impurity stabilities) overnight.

Investigation of Surface Acidity

111

Gauge pressure location

Wafer holder hook (top position)

Glass squared pipe

Wafer holder

To vacuum apparatus

Sample Heater element

IR

KBr windows

Figure 4.5 Schematic representation and photograph of the in situ IR cell. Figure kindly provided by Dr. P. Bazin, LCS, Caen, France

The principle of probe molecule use to investigate a solid surface is conceptually simple; a molecule (whose vibrational modes are well known) in the gas or vapour phase is sent onto the surface to be analysed. The modification of the probe vibrations due to the interaction(s) will reflect the state of the surface itself, i.e. the physico-chemical properties of the solid, the concentration of the adsorption sites, their strength, the mode of coordination, etc. The use and the choice of appropriate probe molecules have been detailed in numerous papers and in particular in reference 28, which should be consulted for further information.

112

Functionalized Inorganic Fluorides

4.3.2

4.3.2.1

Characteristic Reactions and the Detection of Adsorbed Species by a Radiotracer Method State of the Art

A ‘probe reaction’ can be used to make comparative statements about the nature of different acidic surfaces. This is an approach of long standing and there are many applications described in the literature. Two examples will suffice to introduce the topic, since other chapters in this volume deal specifically with reactions that are catalysed by solid metal fluorides and related compounds. The dehydration of isopropanol to give propene has been used in making comparisons of solid acidic oxides2 and the isomerization of 1,2-dibromohexafluoropropane, CBrF2CBrFCF3, to its 2,2-dibromo-isomer, CF3CBr2CF3, is a reaction diagnostic for very strong solid Lewis acids such as ACF.19,23 In the FTIR approach, described above, there is a reasonably direct link from spectroscopic data to relative acidities and, at least for solid oxides, a substantial literature linking spectroscopic properties of probe molecules to acid site strengths. In contrast, the use of probe reactions to make statements about acidity has a smaller literature and a linkage that is less direct. However, providing comparative studies are undertaken under identical or near-identical conditions, the probe reaction approach can provide useful comparative data on catalyst activities and is complementary to the more widely used FTIR method. 4.3.2.2

Anhydrous Hydrogen Chloride as a Probe

The focus here will be on anhydrous HCl, since, (i) it is a strong Brønsted acid (at least in water), (ii) in principle it can behave as a Brønsted acid by dissociative adsorption at a surface that contains oxide (hydroxide) functioning as a Brønsted base and (iii) it could function as a weak Lewis base as a result of associative adsorption at a surface that contains (strong) Lewis acid sites. From this analysis, it would be expected that the behaviour of HCl towards fluorides would be different from its behaviour towards oxides. This comparison is less clear cut however if oxyfluorides and partially hydrolysed (hydroxylated or hydrated) fluoride surfaces are included. Schematic representations of dissociatively adsorbed HCl at medium strong and weak Lewis acid sites of -alumina are shown in Figures 4.6 and 4.7. These are based on spectroscopic observations. The three possible modes of adsorption of anhydrous HCl at a metal fluoride surface, physically adsorbed, associatively adsorbed and dissociatively adsorbed (the latter two cases are chemical adsorption) are represented diagrammatically in Figure 4.8. Anhydrous HCl can be used as a probe either by direct exposure to a surface or it may be generated in situ by a dehydrochlorination reaction (XIII). H CI + HCI (XIII)

Investigation of Surface Acidity

113

CI OH AlTet

OH

HCI

∗ AlTet

AlTet

O

AlTet

+ OH(ads)

O – H2O CI

Al

∗

AlTet O

Figure 4.6 Proposed adsorption of anhydrous HCl at a medium strong Lewis acid site of Z-alumina and the effect of subsequent H2O loss. (Reprinted with permission from A. R. McInroy, D. T. Lundie, J. M. Winfield, C. C. Dudman, P. Jones, S. F. Parker and D. Lennon, Catal. Today, 114, 403–411 Copyright (2006) Elsevier Ltd.)

H O

O ∗ AIOct

AIOct

HCI

AIOct

AIOct

CI

AIOct AIOct

Figure 4.7 Schematic representation of HCl dissociatively adsorbed at a coordinatively unsaturated AlIII site of Z-alumina approximating to the weak Lewis acid type. (Reprinted with permission from A. R. McInroy, D. T. Lundie, J. M. Winfield, C. C. Dudman, P. Jones, S. F. Parker and D. Lennon, Catal. Today, 114, 403–411 Copyright (2006) Elsevier Ltd.)

HCI (g) HCI

HCI

HCI

HCI

HCI

HCI

HCI

phsical adsorption, weakly bound

HCI

HCI

HCI

and/or

CI H

CI

H

associative or dissociative chemical adsorption, strongly bound

Figure 4.8 Three possible modes of adsorption for anhydrous HCl at metal fluorides. (Reprinted with permission from M. Nickkho-Amiry and J. M. Winfield, J. Fluorine Chem., 128, 344–352 Copyright (2007) Elsevier Ltd.)

114

Functionalized Inorganic Fluorides

Dehydrochlorination of hydrochlorocarbons in the vapour phase occurs under homogeneous flow conditions at moderate temperatures.29, 30, 31, 32 Reaction occurs either via a radical chain mechanism, for chlorinated ethanes,29 or via a unimolecular pathway, for t-butyl chloride,30 and for two chloropropanes.31 1,1,1-Trichloroethane undergoes dehydrochlorination by both pathways.32 In the presence of a clean Pyrex surface a, faster, heterogeneous pathway becomes available, for example for ButCl,30 and, more generally, heterogeneous processes, usually observed > 400 K, are the norm for hydrochlorocarbons in the presence of Lewis acidic oxides, such as CH3CCl3 at -alumina,33 CH3CHClCH3 at several acidic oxides and dichloropropane isomers at silica-alumina.34 Heterogeneous, Lewis acid catalysed dehydrochlorination occurs also as one reaction in the complex oxychlorination process catalysed either by CuCl2 supported on -alumina35 or by halide melts.36 In striking contrast to the dehydrochlorination of hydrochlorocarbons at oxides, observed above room temperature, is the room temperature dehydrochlorination of CH3CCl3 in the presence of resublimed solid aluminium trichloride,37 -alumina that has been chlorinated using CCl4 in order to enhance surface Lewis acidity38 or -alumina previously fluorinated at room temperature using sulfur tetrafluoride.39 Dehydrochlorination is accompanied by the formation on the surface of unsaturated oligomers derived from CH2¼CCl2. These are dark red or purple, therefore even a trace reaction can be easily detected. The process is inhibited by traces of water. It should be noted however that not all Lewis acid fluorides give rise to this phenomenon. The behaviour of mixed metal fluorides, Lewis acid composites formed from -CrF3,3H2O and FeF3, resembles that of -alumina rather than the fluorinated -alumina described above.40 Rather surprisingly, ButCl undergoes dehydrochlorination at room temperature in the presence of either SF4-fluorinated -alumina or b-aluminium trifluoride.41 As a result, this species, together with HCl, have been used as probes of surface acidity. The events that are possible at a fluorinated surface following its exposure to ButCl vapour at room temperature are shown in Figure 4.9.

4.3.2.3

Direct Monitoring of Metal Fluoride Surfaces using Chlorine-36 Labelled HCl

Anhydrous HCl can be detected readily in the gas phase above the surface due to its characteristic IR spectrum. By using chlorine-36 labelling ([36Cl] is a b emitting isotope, t1/2 ¼ 3.01  105 y)42 it can be detected as a surface species, physically or chemically adsorbed as indicated in Figure 4.8. Direct radiochemical monitoring of a surface is a very sensitive and well established method, which was developed originally for the study of [14C]-labelled hydrocarbons at supported metal heterogeneous catalysts.43 It has been used extensively in Glasgow for the study of fluoride and fluorinated oxide catalysts.44 The apparatus used is shown in Figure 4.10. The evacuable Pyrex counting cell, which is connected to a gas-handling system, contains two end-window Geiger-Mu¨ller counting tubes, mounted to ensure identical counting geometry, and a moveable Pyrex boat. The use of the apparatus to demonstrate adsorption is illustrated by an experiment in which aliquots of [36Cl]-anhydrous hydrogen chloride are added successively to the cell, which contains a thinly spread layer of the

Investigation of Surface Acidity CH3

H3C

H3C H 3C

115

CH3

CH3 Cl

Cl ∗

H

H3C H3C

+

∗

(C4H8)n F-alumina (SF4)

H+

+

Cl–

H

CxHy species

*

*

H3C

H HCl

H CH3 + other products

β-AlF3

Figure 4.9 Reactions and other events that are possible following the adsorption of ButCl at a metal fluoride surface. Surface adsorbed species are indicated by a broken line to an asterisk. (Reprinted with permission from M. Nickkho-Amiry and J. M. Winfield, J. Fluorine Chem., 128, 344–352 Copyright (2007) Elsevier Ltd.)

benchmark compound, aluminium chlorofluoride, ACF (see Section 4.2.2 above).44 The procedure is as follows. A sample of ACF is dropped into the cell in vacuo via a Pyrex ampoule (A in Figure 4.10), which is attached vertically to the counting cell. Prior to its use, the flamed out ampoule has been loaded with the solid in a glove box. After it is added to the boat, the sample (250–500 mg) is spread as thinly as possible on the base of the boat. The latter is positioned, using a magnet and soft iron bar encased in the boat handle, directly below one of the counters (C1 in Figure 4.10). The arrangement and procedure enable the requirement for an infinitely thin layer of solid to be located directly below one counter, to be approximately fulfilled. The solid and the cell are outgassed thoroughly and then a measured pressure of [36Cl]-labelled anhydrous HCl is expanded into the cell from the gas-handling manifold (B in Figure 4.10). The cell is isolated and its contents allowed to equilibrate for 15 min. Counts from each G.-M. counter are recorded simultaneously on two scaler timers, ideally for a sufficient time to accumulate a substantial number of counts. Since the counting error is the square root of the count number,42 to achieve a relative error of 1% requires 104 counts to be accumulated. This is not always possible and a counting time of 500 s is an acceptable compromise. After counting, the vapour is removed from the cell at room temperature by condensation at 99 K and the sequence: admission of gas, equilibration and counting, repeated. In the experiments described here,

116

Functionalized Inorganic Fluorides

A

B C1

C2

solid counting apparatus

Figure 4.10 A schematic of the Geiger-Mu¨ller direct monitoring process. C1 and C2 are two intercalibrated end window Geiger-Mu¨ller counting tubes, A is an evacuable ampoule, originally containing the solid fluoride; B is an evacuable ampoule from which the labelled volatile probe compound is dispensed. (Reprinted with permission from M. Nickkho-Amiry and J. M. Winfield, J. Fluorine Chem., 128, 344–352 Copyright (2007) Elsevier Ltd.)

usually eight or nine aliquots were used for each series. After removal of the last aliquot, the outputs counts from both counters are recorded. The cell is left evacuated for at least 18 h, usually with intermittent pumping, before additional counts are recorded. Background counts are obtained before and after each experiment and counters are changed when their contamination is suspected. Counters are intercalibrated regularly using a range of pressures of the labelled vapour being used. The intercalibration ratio is the gradient of the linear relationship between counts from the two counters. Experimental data obtained from exposures of H36Cl aliquots to ACF are shown in Figure 4.11. The count from C1, Figure 4.11 (a), reflects the sum of the surface count and that from the cone of gas between the counter end window and the surface. A build up of activity on the surface is indicated, suggesting that as gas and presumably weakly adsorbed species are removed between each addition, some of the H36Cl is strongly bound. The relatively large specific surface area (approximately 100 m2 g1) of ACF is ideal for this technique and results in a relatively high degree of precision. The counts from C2 reflect the H36Cl in the gas phase; their magnitudes are far smaller than those counted by C1 and accordingly the precision is lower. Intercalibration factors are usually close to 1, therefore the count number relationship with C2 data is very similar to the intercalibrated values (C2(ic), shown in Figure 4.11 (b). The surface count is the difference between C1 and C2(ic) and is shown in Figure 4.11 (c). The relation of this variation in surface count across the range of

Investigation of Surface Acidity 4000

13 500 intercalibrated C2 500 s count

(a) 13 000 C1 500 s count

117

12 500 12 000 11 500 11 000

3900

(b)

3800 3700 3600 3500 3400 3300 3200 3100

10 500

0

2

4

6 count No.

8

3000

10

0

2

4

6 count No.

8

10

C1- C2(ic) 500 s surface count

9500 (c) 9000 8500 8000 7500 7000

0

2

4

6 count No.

8

10

Figure 4.11 Sequential exposures of H36Cl aliquots (2.0 kPa) to aluminium chlorofluoride (ACF). (a) C1 counts surface þ the volume of gas directly above; C2 counts an equivalent gas volume; (b) C2(ic) is the intercalibrated count from C2. (c) The derived surface count, obtained from C1 – C2(ic). (Reprinted with permission from M. Nickkho-Amiry and J. M. Winfield, J. Fluorine Chem., 128, 344–352 Copyright (2007) Elsevier Ltd.)

additions, together with the analogous relationship when aliquots of [36Cl]-ButCl vapour are exposed to ACF, are discussed below.

4.4

Experimental Studies of Surface Acidity

The fluoride materials to be highlighted are, firstly, the high surface area (HS) fluorides of aluminium(III)45 and magnesium(II)46 prepared by the sol-gel route and, secondly, derivatives of b-aluminium trifluoride, particularly hydroxy, fluorides prepared by a microwave-assisted solventothermal reaction in aqueous-organic HF media.47,48,49 In both cases the synthetic routes are described in detail in other chapters of this volume. The types of information obtained from FTIR spectroscopy and by using [36Cl]-labelling are described in turn. Finally the acidity of an aluminium hydroxy fluoride having the hexagonal tungsten bronze (HTB) structure,50 which was prepared by the

118

Functionalized Inorganic Fluorides

microwave-promoted solventothermal route referred to above,47 will be described using data from both techniques.

4.4.1

Using FTIR Spectroscopy

Despite the great interest in fluoride compounds for catalysis, their surface properties have been little studied up to now, compared with the corresponding metal oxides. Moreover, the published studies deal mainly with material having specific surface areas less than 20 m2 g1. Recently we have investigated well crystallized and amorphous iron, chromium and aluminum fluorides, all presenting a high specific surface area, up to 300 m2 g1. We characterized first crystalline hexagonal HTB structure fluorides, synthesized in Bordeaux. These solids are constituted by MF6 octahedra linked by corners.50 They can ˚ in be considered as microporous materials, having monodirectional channels of 3–4 A diameter, which can contain water or ammonia. Their thermal stability varies between 423 and 623 K according to the nature of the cation (Fe, Cr, Al or Ga). In addition, we have studied materials having the pyrochlore structures, synthesized in Bordeaux, as well as amorphous compounds of a new type, synthesized using a sol-gel method in Berlin and having a very high specific surface area. The identities of the compounds to be discussed below are summarized in Table 4.1. Table 4.1 Formulation, method of synthesis and specific surface area of the samples synthesized by the ICMCB team in Bordeaux and the HU-Berlin Institut fu¨r Chemie team Composition and structure

Synthesis

HTB-AlF2.2(OH)0.8a HTB AlF3-x(OH)x HS-AlF3 post fluorinated (F2) HS-AlF3 post fluorinated (CFC) Pyrochlore AlF1.8(OH)1.2

nitrate route nitrate (microwave activation) alkoxide route (microwave activation) alkoxide route (sol-gel)

a

alkoxide route (microwave activation)

Specific surface area/m2 g1

Reference

15 82 330

51, 52 47 48

180–310

53 (see also 45) 49

137

Study extended to MF3-x(OH)x , with M ¼ Fe, Cr, Ga.

4.4.1.1

Chemical Impurities and Hydroxyl Groups

Fluorides often present surface impurities as residuals from synthesis.47–53 Their identification can elucidate the reaction mechanisms; in particular, it has been demonstrated by IR spectroscopy that aluminium hydroxy fluorides having the HTB structure contain coordinated or protonated ammonia.47 Curiously, their synthesis involves a nitrate Al(NO3)3,9H2O precursor and ammonia is absent as a reagent. Therefore we have deduced that precursor decomposition implies nitrate reduction to ammonia by isopropanol during the reaction, thus explaining the presence of both ammonia inside the material and acetone

Investigation of Surface Acidity

119

in the filtrate residual. This result has allowed a better understanding of the microwave assisted synthesis.47 In general, the main impurity is residual water, which can generate OH groups. Their localization, inside the bulk or at the surface, is important to understand solid reactivity. A preliminary study of FeIII, AlIII or CrIII HTB compounds51 has shown the presence of a sharp and intense (OH) band, assigned to structural M-OH groups.52 The use of deuterated probe molecules having different sizes has shown that most of these OH groups are not sensitive to H/D exchange; the steric hindrance of the probe molecule is thus greater than the channel size. Therefore hydroxyls are localized inside the channels; their formation is likely to be due to a F/OH substitution induced by water during the synthesis step. This replacement matches well with a fluorine content significantly less than that expected for pure MF3 fluorides, so that it is more appropriate to define such solids as hydroxyfluorides having a MF3x(OH)x formula.51, 52 More recently, a structural study dealing with aluminum based compounds47 has shown that OH groups bridged between two aluminum atoms are localised in particular crystallographic positions, i.e. F1 and F2 sites (Figure 4.12).

Figure 4.12 A view along the c axis of the AlF3 HTB compound. (Reprinted with permission from D. Dambournet, A. Demourgues, C. Martineau, S. Pechev, J. Lhoste, J. Majimel, A. Vimont, J.-C. Lavalley, C. Legein, J.-Y. Buzare´, F. Fayon and A. Tressaud, Chem. Mater., 20, 1459–1469 Copyright (2008) American Chemical Society.)

Infrared spectra of the compounds indicate that the frequency of (OH) bands is relatively low whatever the identity of the cation (AlIII, FeIII, CrIII), compared with those of the corresponding metal oxides, whereas the position of the (OH) mode is unexpectedly high. Analogous variations in (OH) and (OH) wavenumber values have been observed for bridged hydroxyls in zeolites having a small pore size. Thus, the unusual

120

Functionalized Inorganic Fluorides

position for the stretching and bending modes for hydroxyls present in the tunnels of the HTB framework can be explained by the bridged conformation of the hydroxyls in a confined environment.52 For the aluminium compounds we have reported a (OH) component situated at a higher wavenumber (3680 cm1) with respect to that observed for the Al-OH inside the channels (3665 cm1).47,54 This feature is sensitive to the H/D exchange via D2O, as well as to the adsorption of large probe molecules such as lutidine (DMP) or pyridine, therefore we have assigned it to Al-OH groups on the external surface of crystallites; the absence of a confinement effect would be the reason for its higher frequency. This phenomenon is not specific to the HTB structure; the pyrochlore aluminum hydroxy-fluorides (microporous compounds having a related structure) also present a different (OH) frequency between hydroxyls inside the channels (3673 cm1) and those exposed at the crystallite external surface (3720 cm1).49 An unpublished theoretical modeling of structure and surface of partially hydroxylated HTB aluminum fluorides, carried out by the STFC Daresbury Laboratory team, has confirmed such vibrational behaviour. 4.4.1.2

Brønsted Acidity of Metal Fluorides

Brønsted acidity of a series of HTB materials has been studied by probe molecule adsorption (ammonia, pyridine, lutidine, CO) monitored using infrared spectroscopy. We have observed that the results depend on the probe used and on the activation conditions. The OH groups present in the channels are accessible to ammonia only for Fe and Fe/Cr compounds, while Al samples are unaffected by probe substitution.52 ˚ ) is smaller than Structural analysis confirms that the channel size for aluminum (2.42 A ˚ ), corresponding to the observations made using NH3 and consistent with the for iron (2.7 A greater sizes of pyridine or lutidine. Through ammonium cation formation we have shown that Brønsted acidity is correlated with the presence of such OH groups inside the micropores. Another source of acidity would be the presence of HF inside the channels. Unfortunately it was not possible to confirm such an hypothesis by the observation of IR bands corresponding to this species.52 However, the study of hydroxyl band intensity variation versus adsorbed ammonia quantity for the iron and chromium compounds suggests an additional source of Brønsted acidity, because the ammonium species number deduced is greater than the hydroxyl number. Consequently, it would be interesting to adsorb CO on Fe- or Cr-based compounds, because the HF. . . CO complex is claimed to give a characteristic band at 2172 cm1.55 A more complete study of an aluminum HTB compound, stoichiometry AlF2.6(OH)0.4 and a greater specific surface area (82 m2 g1), shows that external OH groups have sufficient acidity to protonate lutidine, but not pyridine.54 Complementary CO adsorption experiments at low temperatures have confirmed the acidic character of those hydroxyls, presenting a remarkable homogeneity in acidity strength, as confirmed by the D(OH) ¼ 3680 ! 3520 cm1 shift, correlated with D(CO) ¼ 160 cm1, corresponding to a (CO) frequency observed at 2173 cm1 (band E in Figure 4.13). This acidity is definitively lower than those for acidic solids such as zeolites, but clearly greater than that reported for OH groups in aluminas. Thus the fluorine effect on the creation of strong Brønsted sites is established.54 Another Brønsted acidity source is the presence of water on the surface of crystalline or amorphous solids. Thus water addition on the HTB compounds transforms coordinated

Investigation of Surface Acidity

121

3525 0.04

0.1 3685

(d)

(g)

(g) (d) 3700 3695 3690 3685 –1 /cm

3580 3540

3500

–1

E 2173

/cm

abs 0.01

(g)

F 2166 D 2183

(a)

Physisorbed species.

B C 2215 2200 A 2235

(f)

2240

2220

2200

2180

2160

2140

2120

2100

–1

wavenumber/cm

2240

2220

2200

2180

2160

2140

2120

2100

–1

wavenumber/cm

Figure 4.13 Left: IR spectra recorded at 100 K after introduction of increasing CO doses at 100 K: 10 mmol g1 (a), 44, 71, 88, 130, 300 mmol g1 (b to f, respectively) then an equilibrium pressure (665 Pa) (g). Right: deconvolution of the n(CO) band envelope on the spectrum (f). Inset: spectral region of the perturbed n(OH) bands. (Reprinted with permission from D. Dambournet, H. Leclerc, A. Vimont, J.-C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, Phys. Chem., Chem. Phys., 11, 1369–1379 Copyright (2009) PCCP Owner Societies.)

ammonia species into ammonium. Similarly, water addition transforms Lewis acid sites into Brønsted sites on activated amorphous compounds. Due to the strong Lewis acidity, water is strongly coordinated on fluorides and generates strong Brønsted sites, as evidenced by CO adsorption.53 On metal oxides, it is well known that water can transform Lewis into Brønsted acidity by dissociative water adsorption on the Mþ – O acid-base pairs, with the consequent creation of M-OH acidic groups. The absence of sufficiently strong basic sites on fluorides, as shown by an unpublished study of propyne adsorption, inhibits dissociative adsorption of water on such materials. 4.4.1.3

Lewis Acidity of Metal Fluorides

The studies performed on a wide range of compounds as indicated in Table 4.1, have allowed three general factors influencing the strength and the number of Lewis acid sites to be distinguished: (a) A cation effect: comparing spectra of adsorbed pyridine on Cr, Fe, Al, Ga HTB compounds, we observe that the strength of the sites varies in the order Al > Ga > Cr  Fe. This order is logical considering the w/r2 ratio, where w is the

122

Functionalized Inorganic Fluorides

electronegativity on the Allred-Rochow scale and r is the ionic radius of the cation. The same order is observed for the thermal stability of the compounds.51 (b) Effect of the amount of fluorine: the frequencies of the  8a and  19b bands for coordinated pyridine are at higher wavenumbers than those observed on the corresponding oxides. It is well known that the electron attracting effect of fluorine increases the acid site strength. This becomes very important, as revealed by the high positions of the  8a and  19b pyridine species at 1628 and 1457 cm1, respectively, approaching the value observed in the py,BF3 complex. This point also explains why the hydroxy fluorides are less acidic than their corresponding fluorides. In particular, the strength of pyrochlore compound (F/Al ¼ 1.8) acid sites is less than those of the HTB compounds (F/Al ¼ 2.6) but greater than those of the corresponding oxides.49 (c) Effect of the structure: when the pores of the material are sufficiently large (as for example in the FeIII HTB compound), ammonia adsorption shows the presence of Lewis acid sites inside the material.52 This increases considerably the number of sites and demonstrates the presence of poorly accessible anionic vacancies. The number of Lewis acid sites at the surface of a range of aluminium fluoride derivatives (quantified through the intensity of the  19b band of strongly coordinated pyridine) has been shown to be related directly to the specific surface area of the compound.48 This finding is particularly relevant when the effectiveness of different Lewis acids, say as catalysts, is being compared and it illustrates very well the importance of the synthetic method used to produce high surface area materials. In many compounds the Lewis acid sites are highly heterogeneous. The use of CO as probe molecule has shown the presence, in addition to the Brønsted type discussed above, of five types of Lewis acid sites on the surface of the AlF2.6(OH)0.4 high surface area HTB compound (Figure 4.13).54 They are labelled A (2235 cm1, D ¼ 94 cm1), B (2215 cm1, D ¼ 74 cm1), C (2200 cm1, D ¼ 59 cm1), D (2183 cm1, D ¼ 42 cm1) and F (2166 cm1, D ¼ 25 cm1). Postfluorination of the sample using CHF3 modifies the numbers of some types of sites; sharp decreases of A, C, D and E sites were observed, while the quantity of B sites was affected to a far smaller extent. To help in the assignment of CO adspecies, a theoretical modeling investigation has been performed. From the morphology of an HTB b-AlF3 crystallite it can be predicted that the exposition of (010) and (001) surfaces is preferential, whereas the (100) surface is unstable.56 Ab initio calculations simulating CO adsorption on the partially hydroxylated, (010) face show that the wavenumber of the coordinated CO species is the highest when the coordinatively unsaturated AlIII pentacoordinated site is surrounded only by fluorine anions. The substitution of F ions by OH groups clearly results in a shift to lower wavenumber. As a consequence, the B site in Figure 4.13 should correspond to AlF5 entities, the C site to a AlF4(OH) environment, while the sites D and F are due to AlF3(OH)2 environments. These assignments agree well with the modifications in band intensities after fluorination. To complete the surface description, the A site will probably correspond to very unstable AlF5 entities on the (100) face, or to Al3þ ions in a tetrahedral coordination. As has been described above, the heterogeneity of the Lewis sites essentially arises from the partially hydroxylated state of many of the aluminum fluorides. A greater

Investigation of Surface Acidity

123

homogeneity of the surface can be obtained via a postfluorination treatment, as a sample thermal treatment under CF3H. The number of strong acid sites so obtained depends on the probe molecule used to measure them (CO: 0.2–0.4 site per nm2; py: 1.2–1.5 site per nm2).45, 48, 53 These results for HTB-AlF2.6(OH)0.4 will be reconsidered below, in the light of complementary radiotracer experiments, to discuss in a complete way the surface behaviour of this fluoride using a variety of methods.

4.4.2 4.4.2.1

Using HCl as a Probe with Detection via [36Cl]-Labelling Aluminium Chlorofluoride

The benchmark compound for this technique is aluminium chlorofluoride (ACF), some of whose properties were described in Section 4.2.2 above. There are many pieces of evidence for this compound being a strong, solid Lewis acid.19, 23 Exposure of ACF to successive aliquots either of anhydrous H36Cl or of [36Cl]-ButCl leads to rather similar behaviour, shown in Figure 4.14. This describes the [36Cl] surface counts resulting from H36Cl (a) or [36Cl]-ButCl additions (b). Although numerical comparisons between (a) and (b) are not possible, since the [36Cl] specific activities of the H36Cl and [36Cl]-ButCl used differ, the very similar qualitative relationship observed suggests that the surface count data observed in (b) are due to adsorbed H36Cl, which results directly from dehydrochlorination of [36Cl]-ButCl at the ACF surface. Removal of volatile material by condensation after count No. 9 in Figure 4.14 (a), leads to a reduction in the surface count, No. 10, and a further reduction after one day, count No. 11. However, the count is still well above background, indicating that a significant fraction of H36Cl is strongly adsorbed. A small increase in the [36Cl] surface count derived from exposures of [36Cl]-ButCl to ACF, count No. 9 in Figure 4.14 (b), is observed after removal of all volatiles, i. e. after count No. 8. This can be accounted for by the assumption that some dehydrochlorination sites with the accompanying adsorption sites for H36Cl are located at grain boundaries within the solid ACF rather than being limited to the exterior surface. The isotope [36Cl], in common with all b emitters, is subject to self-absorption of its radiation.42 It will be detected by the endwindow G. M. counters used (Figure 4.10) only when located at the exterior surface of the sample under investigation. Evidently in the case of ACF, some time must elapse for migration of the [36Cl] species to occur from the bulk to the exterior surface. This phenomenon is observed also for the other compounds described in this section. 4.4.2.2

HS-Aluminium Trifluoride

As explained above HCl can be used directly or as a product from the dehydrochlorination of ButCl. The latter reaction is easily observed over HS-AlF3 at room temperature57 and comparisons between the behaviour of HS-AlF3 with those of fluorinated (by SF4)

-alumina and b-AlF3 (HTB structure) can be made.41,44 They indicate that the greatest reactivity is found for HS-AlF3, either because it has the greatest site density (specific surface areas of samples of this compound as determined by the BET method are usually in the range 200–300 m2 g1)45 or because it has uniquely strong Lewis acid sites. The

124

Functionalized Inorganic Fluorides 10 500 (a)

10 000

500 s surface count

9500 9000 8500 8000 7500 7000 6500 0

2

4

6

8

10

12

count No. 1–9 (1.3 kPa); 10, 11 (after gas removal) 25 000

500 s surface count

24 000

23 000

(b)

22 000

21 000

20 000 0

2

4

6

8

10

count No.1–8(6.7 kPa); 9 (after vapour removal)

Figure 4.14 [36Cl] Surface count relationships from the successive additions of (a) H36Cl and (b) [36Cl]-ButCl to aluminium chlorofluoride (ACF). Line breaks correspond to the removal of the last aliquot of vapour. (Reprinted with permission from M. Nickkho-Amiry and J. M. Winfield, J. Fluorine Chem., 128, 344–352 Copyright (2007) Elsevier Ltd.)

behaviour of H36Cl, either from direct addition or generated in situ, indicates both physical and chemical adsorption.57 In the latter case both exterior (at the gas-solid interface) and interior (boundaries between particles) surfaces are involved. This distinction is possible experimentally due to the self-absorption property of the b radiation emitted from [36Cl] as discussed above for ACF. The exact behaviour observed when [36Cl] surface counts are examined over a range of H36Cl additions depends on the fluorinating agent used to prepare HS-AlF3 from the initial sol-gel. The behaviour of HS-AlF3 samples, which have been prepared either by fluorination with CCl2F2/N2 or by fluorination with aHF/N2 in the second stage of their preparations, towards H36Cl additions is compared in Figure 4.15 (a) and (b).

Investigation of Surface Acidity

125

Exposure of HS-AlF3, which has been prepared by the CCl2F2/N2 route, to successive aliquots of H36Cl, Figure 4.15 (a), leads to a progressive build up of [36Cl] on the surface. Its removal, in vacuo with occasional pumping, is slow enough to be monitored; the initial loss, count Nos. 9–11 is relatively rapid but even after five days, count No. 12, a substantial fraction remains. Evidently much of the H36Cl is strongly bound. The BET area of the material synthesized via the aHF/N2 route is larger, reflected in the larger surface counts shown in Figure 4.15 (b). In this case the [36Cl] surface counts are almost constant over the range of H36Cl additions; the value, count No. 10, measured immediately after removal of the last aliquot, falls to a low level but counts made 24 h later, count Nos 11 and 12, are 10 500

(a)

10 000

500 s surface count

9500 9000 8500 8000 7500 7000

0

28 000

2 4 6 8 10 12 count No. 1–3 (6.0 kPa), 4–8 (8.0 kPa), 9–12 (during gas removal) (b)

26 000

500 s surface count

24 000 22 000 20 000 18 000 16 000 14 000 12 000 10 000 8000 0

2 4 6 8 10 count No. 1–9 (5.3 kPa), 10–12 (after gas removal)

12

Figure 4.15 [36Cl] Surface counts from an high surface area (HS) AlF3 sample, whose precursor was fluorinated with (a) CCl2F2/N2, or (b) with HF/N2, after exposure to sequential aliquots of H36Cl. Line breaks correspond to the removal of the last aliquot of gas. (Reprinted with permission from M. Nickkho-Amiry, G. Eltanany, S. Wuttke, S. Ru¨diger, E. Kemnitz and J.M. Winfield, J. Fluorine Chem., 129, 366–375 Copyright (2008) Elsevier Ltd.)

126

Functionalized Inorganic Fluorides

greater. This appears to reflect diffusion of [36Cl] species from bulk to the exterior surface, as proposed above for ACF. A more extensive comparison of the two types of HS-AlF3 is provided by the dehydrochlorination experiments shown in Figure 4.16 (a) and (b). (a)

10 000

500 s surface count

9000 8000 7000 6000 5000 –2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 count No. 1–8, 10–17 and 19–26 (all 5.3 kPa); 9, 18 and 27 (vapour removed)

18 000

500 s surface count

16 000 14 000 12 000

(b)

10 000 8000 6000 4000 2000 0

–2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 count No. 1–8, 11–19, 21–29 (5.3, 2.7, 8.0 kPa); 9–10, 20, 30–31(after vap. removed)

Figure 4.16 Comparison of the behaviour of [36Cl]-ButCl towards HS-AlF3 samples whose precursors had been fluorinated with (a) CCl2F2/N2 and (b) with HF/N2; three consecutive sets of sequential exposures are shown for each sample. Line breaks correspond to the removal of the last aliquot of vapour in a set and before commencing the next. (Reprinted with permission from M. Nickkho-Amiry, G. Eltanany, S. Wuttke, S. Ru¨diger, E. Kemnitz and J.M. Winfield, J. Fluorine Chem., 129, 366–375 Copyright (2008) Elsevier Ltd.)

Investigation of Surface Acidity

127

In both cases the [36Cl] surface count data are the result of a series of three series of experiments in which eight or nine aliquots of [36Cl]-ButCl were allowed successively to contact a sample of HS-AlF3.57 After each series of additions, the [36Cl] count from the surface in the absence of vapour was recorded, count Nos 9, 18 and 27 in Figure 4.16 (a) and 9–10, 20 and 30–31 in Figure 4.16 (b). Although in neither case is there any substantial evidence for the inhibition of dehydrochlorination by strongly bound H36Cl, it appears that retention of strongly bound H36Cl is relatively less important by HS-AlF3 prepared via the aHF route, Figure 4.16 (b). The difference between the two samples of HS-AlF3 can be rationalized in terms of HF-blocking of the strongest Lewis sites at which, in the absence of HF, HCl would be adsorbed. It appears also that HCl/HF adsorption sites are distinct from sites at which dehydrochlorination occurs.

4.4.2.3

HS-Magnesium Difluoride and Related Materials

Magnesium fluoride is usually regarded as a Lewis base fluoride because in its reactions it is a potential source of F anion. As conventionally prepared, it is not known to be a Lewis acid. It has a moderate specific surface area, 45 m2 g1 and has been used as a support material for heterogeneous catalysts; in this latter context, surface hydroxyl groups generated during manipulation of MgF2 can behave as Brønsted bases.58 However, as prepared by the sol-gel route,46 its behaviour towards H36Cl and [36Cl]-ButCl is rather similar to that of HS-AlF3 as described above. However, it is less active with respect to room temperature dehydrochlorination. For example there is no evidence from FTIR examination of the vapour above HS-MgF2 for HCl formation after ButCl has been admitted, although it is detected easily at the surface by [36Cl] labelling.57 Similar behaviour is shown by the amorphous solid, HS-MgF2/15 mol% FeF3 (equivalent in a crystalline context to a Tanabe solid40,59 in which Lewis acidity is enhanced compared with MF2 by doping with M0 F3). Because of their similar BET areas and pore volumes, [36Cl]-surface counts on 15%FeF3 in HS-MgF2 can be compared with HS-MgF2; the comparison indicates that the 15% FeF3 in HS-MgF2 material is the more effective Lewis acid. A similar finding is indicated from NH3 TPD experiments.57 The behaviour of these two materials with respect to their behaviour towards H36Cl and [36Cl]-ButCl are compared in Figures 4.17 and 4.18.

4.4.2.4

Common Features

The fluoride surfaces whose behaviour has been described above all exhibit the following features when examined by the Geiger-Mu¨ller direct monitoring method: (a) Surface radioactivity is readily detected at each surface following either direct exposure of H36Cl or [36Cl]-ButCl to the thinly spread solid at room temperature. (b) The counts are detectable in the [36Cl]-ButCl case even when the evolution of HCl to the gas phase above the solid is too small to be detected by transmission FTIR. This illustrates the sensitivity of the method but it is aided also by the relatively large BET areas that these solids have.

128

Functionalized Inorganic Fluorides

(c) The [38Cl]-surface activity deposited at the surface is, to a great extent, strongly bound. It is located also within the ‘bulk solid’; this location is not detected by the counters due to self-absorption of the b emitter, [36Cl] but its existence is inferred by the surface count behaviour following removal of radioactive vapour and weakly bound [36Cl] species. Possible ways of formulating the adsorbed states of H36Cl are discussed below (in Section 4.4.4). 26 000 (a)

500 s surface count

24 000 22 000 20 000 18 000 16 000 14 000 12 000 10 000 0

2 4 6 8 10 count No. 1–8 (3.3 kPa), 9,10 after gas removal

95 000

500 s surface count

90 000 (b) 85 000 80 000 75 000 70 000 0

2 4 6 8 10 count No. 1–8 (2.7 kPa), 9,10 (after gas removal)

Figure 4.17 Comparisons among [36Cl] surface counts from H36Cl aliquots in contact with (a) high surface area (HS) MgF2 and (b) 15 mol% HS-FeF3 in HS-MgF2. Line breaks correspond to the removal of the last aliquot of gas. (Reprinted with permission from M. Nickkho-Amiry, G. Eltanany, S. Wuttke, S. Ru¨diger, E. Kemnitz and J.M. Winfield, J. Fluorine Chem., 129, 366–375 Copyright (2008) Elsevier Ltd.)

Investigation of Surface Acidity

129

7000 (a) 500 s surface count

6500 6000 5500 5000 4500 4000 3500 0

2

4

6

8

10

count 1–8 (4.0 kPa), 9,10 (after vapour removal)

14 000 (b)

500 s surface count

13 000 12 000 11 000 10 000 9000 8000 7000 0

2

4

6

8

10

count No. 1–8 (4.0 kPa), 9 (after vapour removed)

Figure 4.18 Comparisons between [36Cl] surface counts from [36Cl]-ButCl aliquots in contact with (a) HS-MgF2 and (b) 15 mol% HS-FeF3 in HS-MgF2. Line breaks correspond to the removal of the last aliquot of vapour. (Reprinted with permission from M. Nickkho-Amiry, G. Eltanany, S. Wuttke, S. Ru¨diger, E. Kemnitz and J.M. Winfield, J. Fluorine Chem., 129, 366–375 Copyright (2008) Elsevier Ltd.)

4.4.3

Metal Fluoride Surfaces that Contain Surface Hydroxyl Groups: Aluminium Hydroxy Fluorides with the Hexagonal Tungsten Bronze Structure

The focus in this section is on derivatives of the metastable b phase of aluminium trifluoride, which has the hexagonal tungsten bronze structure with AlF6 octahedra linked

Functionalized Inorganic Fluorides

130

to form hexagonal hollow tubes running through the lattice50 (see also Figure 4.12). In addition to its use as a laboratory heterogeneous catalyst,1 it can be considered to be a relative of amorphous HS-AlF3; this is illustrated, for example, by a recent comparative study of vibrational spectra of AlF3 forms.60 b-Aluminium trifluoride was one of the earliest fluoride Lewis acid surfaces to be investigated by the use of py as a probe species for surface acidity.61 Because of the way in which it is prepared, b-AlF3 usually has water located in its hexagonal channels.50 Water can migrate to the surface where it can coordinate to Lewis acid AlIII sites or initiate hydrolysis to form hydroxylated AlIII sites. Both types of site can be probed using H36Cl;41,62 the steps that are believed to be involved are shown schematically in Figure 4.19. Al

Al F

F H2O

Al

Al

and F

F

F Al

OH

Al

then 523 K in vacuo

F

F

OH2

723 K/He

F

Al

F

major

F minor

hydrated site H

+ HF

Al

hydroxylated site

H room temp.

O Al F

>373 K

OH2.ClH

+ HCl

Al F

F

+ H2O + HCl

or pump

Al F

F

F

HCl at room temperature Cl

OH room temp. + HCl

Al F

Al

to 623 K

F

F

+ H2O F

chlorination of hydroxyl sites

Figure 4.19 Schematic describing the proposed formation of hydrated and hydroxylated surface sites at HTB b-AlF3 and the subsequent adsorption/desorption of anhydrous HCl. (Reprinted with permission from C. H. Barclay, H. Bozorgzadeh, E. Kemnitz, M. NickkhoAmiry, D. E. M. Ross, T. Skapin, J. Thomson, G. Webb and J. M. Winfield, J. Chem. Soc., Dalton Trans., 40–47 Copyright (2002) Royal Society of Chemistry.)

It follows from these proposals that water coordinated to a surface AlIII site can give rise to Brønsted activity and on exposure to HCl behaves as a Brønsted base. The replacement of surface –OH groups by Cl, which is a result from HCl treatment above room temperature,62 is reminiscent of the chlorination of alumina by HCl38, 63 (cf. Figures 4.6 and 4.7).

Investigation of Surface Acidity

131

A more detailed examination becomes possible for compounds where hydroxyl groups are deliberately incorporated into the structure as opposed to arising from trace hydrolysis. Here we remind the reader of several points that were made above in Section 4.4.1. In the HTB solids of stoichiometry, MF3x(OH)x, where M is Al, Cr or Fe,51 both Brønsted and Lewis acidity can be demonstrated by FTIR spectroscopy using the basic probe molecules, NH3 and py.52 Even more information is available if the synthesis route is modified to produce solids having greater surface areas, for example by use of the microwave-activated, solventothermal process.47 The HTB-structured solid whose stoichiometry is AlF2.6(OH)0.4 is a particularly good example and its surface acidity has been studied using pyridine, lutidine (2,6-dimethylpridine), CO, H36Cl and [36Cl]-ButCl as surface probes.54 The pertinent FTIR spectroscopic data for the b-AlF2.6(OH)0.4 surface, which were discussed in Section 4.4.1, are summarised for convenience in Table 4.2. Making comparisons with related studies on aluminas,6, 9, 64 the presence of strong surface Lewis acidity is evident and there is good evidence for Brønsted acidity associated with the –OH groups.

Table 4.2 Characterization by FTIR spectroscopy of the Lewis/Brønsted acidic sites displayed by the b-AlF2.6(OH)0.4 surface. Data taken from reference 54. (Reproduced by permission of the PCCP Owner Societies.) Probe IR bands

Lutidine  8a (CC)/cm1 Pyridine  8a/cm1 CO (CO)/cm1

Bands due to Lewis acidity

1615–1620 1610

1628 1620

Bands due to Brønsted acidity

1652, 1631a

–

Proton affinity/kJ mol1

963

a

912

2235 2220–2215 2200 2183 2166 2173 D(OH) 160 598

Band is v8b.

The use of several probe molecules has enabled complementary information to be obtained, the detail of which depends on the identity of the probe. In particular, heterogeneity in the Lewis site strength is indicated on the exterior surface, as demonstrated by the frequency range and the basicity displayed by a probe molecule.54 Using py as the probe, two types of Lewis acid sites are detected: Figure 4.20; the first one, denoted L1 in Figure 4.20, is very strong ( 8a py band at 1628 cm1) but is present only in low concentration on the surface (approximately 0.2 sites nm2). The second type, denoted L2, is weaker ( 8a py band at 1620–1623 cm1) but is more abundant (approximately 1.2 sites nm2). Adsorption of CO, which was illustrated in Figure 4.13, indicates the existence of five different types of Lewis site (A-D and F) and a Brønsted site (E), the latter being responsible for the shift induced in the (OH) band (Figure 4.13, inset A). Pyridine-CO co-adsorption experiments reveal the relationships among the two sets of individual

Functionalized Inorganic Fluorides

132

A

L2

3680

B

Abs 0.01 L1 (c-b)

0.05 3500

3000

2500

/cm–1 1635

1625

1615 /cm–1

3680 1454

0.5

(a)

Abs 0.04

1620

1600

(b) (c) (d) (e) (f)

1628

1457

(g) 3680

3640 /cm–1

3600

1650

1600

1550 /cm–1

1500

1450

Figure 4.20 IR spectra of b-AlF2.6(OH)0.4 after activation at 573K before (a) and after introduction of an equilibrium pressure (133 Pa) of pyridine (b-g); (b) evacuation at room temperature under vacuum and thermodesorption at (c) 323, (d) 423, (e) 473, (f) 523 and (g) 573 K. Inset A: difference IR spectra after introduction of an equilibrium pressure (133 Pa) of pyridine followed by evacuation at r.t. Inset B: deconvolution (dotted lines) of the n8a vibrational mode (at room temperature). (Reprinted with permission from D. Dambournet, H. Leclerc, A. Vimont, J.-C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, Phys. Chem. Chem. Phys. 11, 1369–1379 Copyright (2009) PCCP Owner Societies.)

experiments. The L1 and L2 types identified from py adsorption/desorption, Figure 4.20, are both heterogeneous; L1 and L2 sites are each related to two sites. Site A and site B, identified by (CO) bands at 2235 and 2215 cm1, are included in the L1 envelope; site C and site D, characterized by (CO) bands at 2200 and 2183 cm1 are related to L2. Adsorption of CO reveals also the presence of weaker Lewis acid sites characterized by a (CO) band at 2166 cm1 (site F in Figure 4.13). The different types, and strengths, of the Lewis sites are suggested to arise as a consequence of the various F/OH anionic environments that exist, presumed to be related to those revealed previously in the bulk by 27Al NMR.47 The IR spectrum of lutidine (DMP) adsorbed on b-AlF2.6(OH)0.4 and displayed in Figure 4.21, shows the presence of coordinated lutidine ( 8a at 1615,  8b at 1590 cm1) and lutidinium species ( 8a at 1652 and  8b at 1630 cm1). The latter bands indicate that surface Brønsted acid sites are present. In agreement with this

Investigation of Surface Acidity

133

conclusion, subtracted spectra in the hydroxyl stretching region, suffer a decrease in the intensity of the 3680 cm1 band (Figure 4.21, inset). Despite the steric hindrance induced by the methyl groups of lutidine, it appears that accessible acidic OH groups are present at the surface. Abs 0.2 (a) (b)

3680

(b)-(a) 3680

3640 /cm–1

3600

Physisorbed Abs 0.1

Lewis Brønsted 1615 1652 1631 Adsorption Evacuation RT 1620

373 K 423 K

1700

1650 1600 Wavenumber/cm–1

1550

Figure 4.21 Difference IR spectra of coordinated lutidine on b-AlF2.6(OH)0.4 after adsorption of 133 Pa at equilibrium pressure, followed by desorption under vacuum at room temperature (293 K), 373 and 423 K. The dotted curve is the difference curve between desorption obtained at 373 and 473 K. Inset: spectra in the n(OH) region of the sample activated at 573 K (a) and after adsorption of lutidine (b). (Reprinted with permission from D. Dambournet, H. Leclerc, A. Vimont, J.-C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, Phys. Chem. Chem. Phys. 11, 1369–1379 Copyright (2009) PCCP Owner Societies.)

To summarize, experiments with CO and lutidine, described above, both indicate that surface OH groups ((OH) at 3680 cm1) are present, which generate homogeneous Brønsted acid sites with a medium acid strength (D(OH) ¼ 160 cm1, (CO) ¼ 2173 cm1), able to protonate lutidine but not pyridine.

134

Functionalized Inorganic Fluorides

The behaviour of b-AlF2.6(OH)0.4 towards H36Cl is shown in Figure 4.22.54 From this sample, loosely bound H2O, including as much as possible of the water originally located in the hexagonal channels, had been removed before exposure to H36Cl. The interaction of H36Cl is progressive with increasing number of aliquots used and both surface and bulk appear to be involved. Most obviously, the interaction should involve Al-OH groups both those located at the exterior surface and in the hexagonal channels. Migration of H36Cl appears to occur during the storage period in vacuo, (from count No. 9 to count No. 10 in Figure 4.22) from hexagonal channels in the HTB-structure, where [36Cl] b radiation cannot be detected because of self absorption42 of [36Cl], to sites on the exterior surface where it is detectable.

13 000

500 s surf. count

12 000 11 000 10 000 9000 8000 7000 6000 0

2 4 6 8 10 count No. 1–8 (3.3 kPa), 9,10 after gas removal

Figure 4.22 Surface counts from anhydrous b-AlF2.6(OH)0.4 during exposure to H36Cl. The line break corresponds to the removal of gas. (Reprinted with permission from D. Dambournet, H. Leclerc, A. Vimont, J.-C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, Phys. Chem. Chem. Phys. 11, 1369–1379 Copyright (2009) PCCP Owner Societies.)

The activity of b-AlF2.6(OH)0.4 with respect to room temperature dehydrochlorination of [36Cl]-ButCl is significantly smaller than that observed for HS-AlF3 or even, possibly, for HS-MgF2. Deposition of [36Cl]-species on the surface, although observed, could not be quantified.54 However, deposition of H36Cl could be achieved using a sample of b-AlF2.6(OH)0.4 from which water had not been exhaustively removed, b-AlF2.6(OH)0.4,xH2O, x is approximately 0.12. It appears therefore that some Lewis acid sites are present even in the presence of the base H2O. The apparent promotional role of H2O in the room temperature dehydrochlorination of ButCl may be the result of migration of the HCl liberated to a nearby water molecule, where it will be strongly adsorbed (cf. Figure 4.19). The presence of hydroxyl groups introduces an extra dimension to the chemistry as OH groups, both located at the surface and in the channels, can behave as Brønsted acids,

Investigation of Surface Acidity

135

towards CO or lutidine and as Brønsted bases towards the potentially strong Brønsted acid, HCl. Although less reactive with respect to dehydrochlorination than, for example HS-AlF3, retention of H36Cl is substantial. The obvious comparison is with the behaviour of H36Cl at conventionally prepared b-AlF3, where the slow desorption of H2O.HCl has been observed at room temperature (Figure 4.19).41 The behaviour of the HTB hydroxy fluorides also provides a bridge to the behaviour of partially hydrolysed metal fluoride surfaces, to be discussed below.

4.4.4

Possible Geometries for HCl Adsorbed at Metal Fluoride Surfaces: Relation to Oxide and Oxyfluoride Surfaces

Throughout this account of surface acidity we have attempted to emphasize the relationships among oxides and fluorides in the context of acidic surfaces. The development of high surface-area metal fluorides makes this comparison far easier than has previously been possible. The various types of interaction between such surfaces and anhydrous HCl are cases in point. This aspect was introduced in Section 4.3.2.2 above and possible modes of adsorption of HCl at metal fluoride surfaces were sketched in Figure 4.8. Anhydrous HCl is commonly dissociatively adsorbed at aluminas. For example, analysis of vibrational data for HCl adsorbed at -alumina63 led to the pictorial representations of HCl adsorbed at medium strong and at a site approximating to a weak Lewis acid type (cf. Figure 4.1) shown in Figures 4.6 and 4.7 respectively. In addition there is vibrational spectroscopic evidence for associative adsorption, molecular HCl being either physisorbed or weakly chemisorbed; in either event the species is desorbed completely by 423 K.63 Dissociative adsorption is also a likely outcome when HCl is exposed to high surface area oxyfluorides. This has been studied at room temperature using H36Cl and the direct G. M. monitoring technique (Figure 4.10) with a series of fluorinated chromia and lightly-doped MgII and ZnII chromia aerogels.65 Significant ( 60–97%) fractions of the H36Cl aliquots were retained by the surfaces, the behaviour being comparable to that observed towards

-alumina. Associative adsorption is a possibility at the surfaces of binary fluorides such as HS-AlF3 (Section 4.4.2.2) and HS-MgF2 and its derivatives (Section 4.4.2.3). These situations have not yet been studied by vibrational spectroscopy, so the inference is made solely on the basis of [36Cl] experiments. Associative adsorption requires HCl to behave as a Lewis base and this type of behaviour is expected only if very strong Lewis acid sites are present.57 However, there are two other possible explanations for the significant degrees of H36Cl retention that are observed. Firstly, it is possible to envisage hydrogen bonding between HCl and surface fluoride and secondly, and perhaps more plausibly, HCl could interact with surface hydrated or hydroxylated sites that are the result of hydrolysis to give rise to a dissociatively adsorbed species. This situation is envisaged to be relevant to the HTB solids b-AlF2.6(OH)0.4 and b-AlF2.6(OH)0.4.xH2O, as some of the Lewis acid sites in b-AlF2.6(OH)0.4 will have mixed F/OH nearest neighbours.54 The possible adsorbed states for HCl at a hypothetical fluoride/hydroxide surface, where the solid is comprised of an aggregate of nanoparticles, is shown diagrammatically in Figure 4.23. Although the picture is speculative, the composite is consistent with all the observations on these solids made to date.

136

Functionalized Inorganic Fluorides F Al

Al F

OH Al

F

F

Al O

F

HCl

ClH

F Al

Al HCl

OH2 Al

F

HCl

Cl

F

HCl

Al O

F

HCl exterior surface

HCl

HCl

interior surface HCl

Figure 4.23 Schematic of HCl adsorption at an hypothetical fluoride-hydroxide surface with the bulk having an aggregated nanoparticle structure

4.5

Conclusions

In this chapter recent developments in the study of the acidity in metal fluorides have been surveyed. The relatively large specific surface areas encountered aid the study of acidic metal fluorides both by FTIR spectroscopy and by a reaction-specific radiotracer approach. Where the structure of the solid is known, the enhanced surface area and the application of both techniques enables a relatively detailed picture of the acidic properties to be obtained. This is the situation for the HTB-aluminium hydroxy fluoride, Figure 4.24. FTIR studies with a range of Lewis base probe molecules point to the presence of a range of Lewis acid sites in HS-metal fluorides, some of which are of higher acidity than in the corresponding oxides. The application of a radiotracer method in which [36Cl] labelled species are used to track species formed at the surface, is less direct than the FTIR method, as structural information is not possible. However this approach does enable features of the bulk to be related to surface events. Because it is also amenable to the use of vacuum and dry box techniques, it is very useful for the study of hydrolytically sensitive fluoride surfaces. In many respects therefore, the radiotracer and FTIR approaches to the examination of surface acidity are complementary. Information that relates to chemical speciation cannot

Investigation of Surface Acidity = F/OH Al

137

Me

Al

N

Me Me

Cl Al

Al

ClH N

Al

Al

CO

Figure 4.24 A schematic of the investigation of HTB AlF2.6(OH)0.4 using adsorption of the probe molecules CO, pyridine, lutidine and [36Cl]-hydrogen chloride. (Reprinted with permission from D. Dambournet, H. Leclerc, A. Vimont, J.-C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, Phys. Chem. Chem. Phys. 11, 1369–1379 Copyright (2009) PCCP Owner Societies.)

be obtained directly from radiochemical monitoring but is directly available by FTIR. The use of a b emitter enables the fate of an adsorbed species to be ‘tracked’ in favourable circumstances from the bulk to the surface.

References [1] A. Moerkerken, B. Behr, M. A. Noordeloos-Maas and C. Boelhouwer, J. Catal., 24, 177–180 (1972); H. J. Reitsma and C. Boelhouwer, J. Catal, 33, 39–46 (1974); E. Kemnitz and J. M. Winfield, Advanced Inorganic Fluorides: Synthesis, Characterization and Applications, T. Nakajima, B. Zˇemva and A. Tressaud (Eds) Elsevier Science SA, Amsterdam, 2000, ch. 12, 367–401. [2] A. S. C. Brown and J. S. J. Hargreaves, Green Chem., 1, 17–20 (1999). [3] S. A. Lermontov, L. L. Ushakova and N. V. Kuryleva, J. Fluorine Chem., 129, 332–334 (2008). [4] A. Majid, S. Argue, D. Kingston and S. Lang, J. Fluorine Chem., 128, 1012–1018 (2007). [5] S. Chaudhuri, P. Chupas, B. J. Morgan, P. A. Madden and C. P. Grey, Phys. Chem. Chem. Phys., 8, 5045–5055 (2006). [6] G. Busca, Phys. Chem. Chem. Phys., 1, 723–736 (1999). [7] C. Morterra G. Magnacca and V. Bolis, Catal. Today, 70, 43–58 (2001). [8] D. T. Lundie, A. R. McInroy, R. Marshall, J. M. Winfield, P. Jones, C. C. Dudman, S. F. Parker, C. Mitchell and D. Lennon, J. Phys. Chem., B, 109, 11592–11601 (2005). [9] H. Kno¨zinger, Adv. Catal., 25, 184–271 (1976); C. Morterra, A. Chiorino, G. Ghiotti and E. Garrone, J. Chem.Soc., Faraday Trans. 1, 75, 271–288 (1979). [10] A. Travert, A. Vimont, A. Sahibed-Dine, M. Daturi, J.–C. Lavalley, Appl. Catal. A: General, 307, 98–107 (2006). [11] A. Vimont, J.-C. Lavalley, A. Sahibed-Dine, C. Otero Arean, M. Rodriguez Delgado, M. Daturi, J. Phys. Chem. B, 109, 9656–9664 (2005). [12] A. Corma, P. Botella and C. Mitchell, Chem. Commun., 2008–2010 (2004). [13] C. Braddock, Green Chem., 3, G26–G32 (2001); M. Shi and S.-C. Cui, Chem. Commun., 994–995 (2002); M. Shi and S.-C. Cui, J. Fluorine Chem., 113, 207–209 (2002). [14] A. G. M. Barrett, D. C. Braddock, D. Catterick, D. Chadwick, J. P. Henschke and R. M. McKinnell, Synlett, 847–849, (2000); X. Hao, A. Yoshida and N. Hoshi, J. Fluorine Chem., 128, 1396–1401 (2007).

138

Functionalized Inorganic Fluorides

[15] X. Hao, A. Yoshida and J. Nishikido, Green Chem., 6, 566–569 (2004); X. Hao, A. Yoshida and J. Nishikido, J. Fluorine Chem., 127, 193–199 (2006); W.-B. Yi, C. Cai and X. Wang, J. Fluorine Chem., 128, 919–924 (2007); W.-B. Yi and C. Cai, J. Fluorine Chem., 129, 524–528 (2008); M.-G. Shen, C. Cai and W.-B. Yi, J. Fluorine Chem., 129, 541–544 (2008). [16] R. J. Gillespie, T. E. Peel and E. A. Robinson, J. Am. Chem. Soc., 93, 5083–5087 (1971). [17] J. Sommer, S. Schwartz, P. Rimmelin and P. Canivet, J. Am. Chem. Soc., 100, 2576–2577 (1978); D. Fa˘rcas¸ iu, S. L. Fisk, M. T. Melchior and K. D. Rose, J. Org. Chem., 47, 453–457 (1982); D. Fa˘rcas¸ iu, Acc. Chem. Res., 15, 46–51 (1982). [18] T. A. O’Donnell, Advanced Inorganic Fluorides: Synthesis, Characterization and Applications, T. Nakajima, B. Zˇemva and A. Tressaud (Eds.) Elsevier Science SA, Amsterdam, 2000, ch. 11, 329–366. [19] T. Krahl and E. Kemnitz, J. Fluorine Chem., 127 663–678 (2006). [20] K. O. Christe, D. A. Dixon, D. McLemore, W. W. Wilson, J. A. Sheehy and J. A. Boatz, J. Fluorine Chem., 101, 151–153 (2000). [21] H. D. B. Jenkins, H. K. Roobottom and J. Passmore, Inorg. Chem., 42, 2886–2893 (2003). [22] H. D. B. Jenkins, I. Krossing, J. Passmore and I. Raabe, J. Fluorine Chem., 125, 1585–1592 (2004). [23] T. Krahl, R. Sto¨sser, E. Kemnitz, G. Scholz, M. Feist, G. Silly and J.-Y. Buzare´, Inorg. Chem., 42, 6474–6483 (2003). [24] V. A. Petrov and C. G. Krespan, J. Fluorine Chem., 102, 199–204 (2000) [25] G. G. Belen’kii, V. A. Petrov and P. R. Resnick, J. Fluorine Chem., 108, 15–20 (2001). [26] C. G. Krespan and D. A. Dixon, J. Fluorine Chem., 77, 117–126 (1996); A. G. Sievert, C. G. Krespan and F. J. Weigert (to DuPont), US Patent 5,157,171 (1992). [27] T. Krahl and E. Kemnitz, Angew. Chem., Int. Ed., 43, 6653–6656 (2004). [28] E. Payen, J. Grimblot, J.-C. Lavalley, M. Daturi and F. Mauge´, Application of Vibrational Spectroscopy in the Characterization of Oxides and Sulfides Catalysts, in Handbook of Vibrational Spectroscopy, ed. J. M. Chalmers and P. R. Griffith, John Wiley & Sons, Inc., New York, 2002, vol. 4. [29] D. H. R. Barton, J. Chem. Soc., 148–155 (1949); D. H. R. Barton and K. E. Howlett, J. Chem. Soc., 155–164 and 165–169 (1949); D. H. R. Barton and K. E. Howlett, J. Chem. Soc., 2033–2038 (1951). [30] D. H. R. Barton and P. F. Onyon, Trans. Farad. Soc., 45 725–735 (1949). [31] D. H. R. Barton and A. J. Head, Trans. Farad. Soc., 46, 114–122 (1950). [32] D. H. R. Barton and P. F. Onyon, J. Am. Chem. Soc., 72, 988–995 (1950). [33] T. H. Ballinger and J. T. Yates, Jr.,J. Phys. Chem., 96, 1417–1423 (1992); T. H. Ballinger, R. S. Smith, S. D. Colson and J. T. Yates, Jr., Langmuir, 8, 2473–2478 (1992). [34] C. Pistarino, E. Finocchio, G. Romezzano, F. Brichese, R. Di Felice and G. Busca, Ind. Eng. Chem. Res., 39, 2752–2760 (2000); C. Pistarino, E. Finocchio, M. A. Larrubia, B. Serra, S. Braggio, G. Busca and M. Baldi, Ind. Eng. Chem. Res., 40, 3262–3269 (2001). [35] D. Carmello, E. Finocchio, A. Marsella, B. Cremaschi, G. Leofanti, M. Padovan and G. Busca, J. Catal., 191, 354–363 ((2000). [36] C. N. Kenney, Catal. Rev. – Sci. Eng., 11, 197–224 (1975). [37] D. G. McBeth, J. M. Winfield, B. W. Cook and N. Winterton, J. Chem. Soc., Dalton Trans., 671–676 (1990). [38] J. Thomson, G. Webb and J. M. Winfield, J. Mol. Catal., 67, 117–126 (1991); J. Mol. Catal., 68, 347–354 (1991). [39] J. Thomson, G. Webb, J. M. Winfield, D. W. Bonniface, C. Shortman and N. Winterton, Appl. Catal. A, 97, 67–76 (1993); A. Bendada, D. W. Bonniface, F. McMonagle, R. Marshall, C. Shortman, R. R. Spence, J. Thomson, G. Webb, J. M. Winfield and N. Winterton, Chem. Commun., 1947–1948 (1996). [40] B. Adamczyk, A. Hess and E. Kemnitz, J. Mater. Chem., 6, 1731–1735 (1996). [41] C. H. Barclay, H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, D. E. M. Ross, T. Skapin, J. Thomson, G. Webb and J. M. Winfield, J. Chem. Soc., Dalton Trans., 40–47 (2002). [42] W. D. Ehmann and D. E. Vance, Radiochemistry and Nuclear Methods of Analysis, John Wiley & Sons, Inc., New York (1991).

Investigation of Surface Acidity

139

[43] A. S. Al-Ammar and G. Webb, J. Chem. Soc., Faraday Trans., 174, 195–205 (1978). [44] M. Nickkho-Amiry and J. M. Winfield, J. Fluorine Chem., 128, 344–352 (2007). [45] E. Kemnitz, U. Grob, St. Ru¨diger and S. Chandra Shekar, Angew. Chem., Int. Ed., 42, 4251–4254 (2003); St. Ru¨diger, U. Grob, M. Feist, H. A. Prescott, S. Chandra Shekar, S. I. Troyanov and E. Kemnitz, J. Mater. Chem., 15, 588–597 (2005); St. Ru¨diger, G. Eltanany, U. Grob and E. Kemnitz, J. Sol-Gel Sci. Technol., 41, 299–311 (2007); St. Ru¨diger, U. Grob and E. Kemnitz, J. Fluorine Chem., 128, 353–368 (2007). [46] J. Krishna Murthy, U. Grob, St. Ru¨diger, E. Kemnitz and J. M. Winfield, J. Solid-State Chem., 179, 739–746 (2006); St. Wuttke, G. Scholz, St. Ru¨diger and E. Kemnitz, J. Mater. Chem., 17, 4980–4988 (2007). [47] D. Dambournet, A. Demourgues, C. Martineau, S. Pechev, J. Lhoste, J. Majimel, A. Vimont, J.-C. Lavalley, C. Legein, J.-Y. Buzare´, F. Fayon and A. Tressaud, Chem. Mater., 20, 1459–1469 (2008). [48] D. Dambournet, G. Eltanany, A. Vimont, J.-C. Lavalley, J.-M. Goupil, A. Demourgues, E. Durand, J. Majimel, St. Ru¨diger, E. Kemnitz, J.M. Winfield and A. Tressaud, Chem. Eur. J., 14, 6205–6212 (2008). [49] D. Dambournet, A. Demourgues, C. Martineau, E. Durand, J. Majimel, A. Vimont, H. Leclerc, J.-C. Lavalley, M. Daturi, C. Legein, J.-Y. Buzare´, F. Fayon, A. Tressaud, J. Mater. Chem., 18, 2483–2492 (2008). [50] A. Le Bail, C. Jacoboni, M. Leblanc, R. De Pape, H. Duroy and J. L. Fourquet, J. Solid State Chem., 77, 96–101 (1988); see also, N. Herron, D. L. Thorn, R. L. Harlow, G. A. Jones, J. B. Parise, J. A. Fernandez-Baca and T. Vogt, Chem. Mater., 7, 75–83 (1995); C. Alonso, A. Morato, F. Medina, F. Guirado, Y. Cesteros, P. Salagre, J. E. Sueiras, R. Terrado and A. Giralt, Chem. Mater., 12, 1148–1155 (2000). [51] L. Francke, E. Durand, A. Demourgues, A. Vimont, M. Daturi and A. Tressaud, J. Mat. Chem., 13, 2330–2340 (2003). [52] A. Vimont, J.-C. Lavalley, L. Francke, A. Demourgues, A. Tressaud and M. Daturi, J. Phys. Chem. B, 108, 3246–3255 (2004). [53] T. Krahl, A. Vimont, G. Eltanany, M. Daturi and E. Kemnitz, J. Phys. Chem. C, 111, 18317–18325 (2007). [54] D. Dambournet, H. Leclerc, A. Vimont, J.-C. Lavalley, M. Nickkho-Amiry, M. Daturi and J. M. Winfield, Phys. Chem. Chem. Phys., 11, 1369–1379 (2009). [55] L. Andrews and R.D. Hunt, J. Phys. Chem., 92, 81–85 (1988). [56] A. Wander, C. L. Bailey, S. Mukhopadhyay, B. G. Searle, N. M. Harrison, J. Phys. Chem. C, 112, 6515–6519 (2008). [57] M. Nickkho-Amiry, G. Eltanany, St. Wuttke, St. Ru¨diger, E. Kemnitz and J. M. Winfield, J. Fluorine Chem., 129, 366–375 (2008). [58] M. Wojciechowska, M. Zielin´ski and M. Pietrowski, J. Fluorine Chem., 120, 1–11 (2003). [59] K. Tanabe, T. Sumiyoshi, K. Shibata, T. Kiyoura and J. Kitagawa, Bull. Chem. Soc. Jpn., 47, 1064–1066 (1974); E. Kemnitz, Y. Zhu and B., Adamczyk, J. Fluorine Chem., 114, 163–170 ¨ nveren and E. Kemnitz, J. Fluorine (2002); J. Krishna Murthy, U. Grob, St. Ru¨diger, E. U Chem., 125, 937–949 (2004). [60] U. Gross, St. Ru¨diger, E. Kemnitz, K.-W. Brzezinka, S. Mukhopadhyay, C. Bailey, A. Wander and N. Harrison, J. Phys. Chem. A, 111, 5813–5819 (2007). [61] A. Hess and E. Kemnitz, J. Catal., 149, 449–457 (1994). [62] H. Berndt, H. Bozorg Zadeh, E. Kemnitz, M. Nickkho-Amiry, M. Pohl, T. Skapin and J. M. Winfield, J. Mater. Chem., 12, 3499–3507 (2002). [63] A. R. McInroy, D. T. Lundie, J. M. Winfield, C. C. Dudman, P. Jones, S. F. Parker and D. Lennon, Catal. Today, 114, 403–411 (2006). [64] C. Morterra, S. Coluccia, E. Garrone and G. Ghiotti, J. Chem. Soc., Faraday Trans. 1. 75, 289–304 (1979). [65] H. Bozorgzadeh, E. Kemnitz, M. Nickkho-Amiry, T. Skapin and J. M. Winfield, J. Fluorine Chem., 121, 83–92 (2003).

5 Probing Short and Medium Range Order in Al-based Fluorides using High Resolution Solid State Nuclear Magnetic Resonance and Parameter Modelling Christophe Legein1, Monique Body2, Jean-Yves Buzare´2, Charlotte Martineau1,3 and Gilles Silly4 1

Laboratoire des Oxydes et Fluorures, CNRS UMR 6010, Institut de Recherche en Inge´nierie Mole´culaire et Mate´riaux Fonctionnels, CNRS FR 2575,Universite´ du Maine, Avenue Olivier Messiaen, 72085 Le Mans Cedex 9, France 2 Laboratoire de Physique de l’Etat Condense´, CNRS UMR 6087, Institut de Recherche en Inge´nierie Mole´culaire et Mate´riaux Fonctionnels, CNRS FR 2575, Universite´ du Maine, Avenue Olivier Messiaen, 72085 Le Mans Cedex 9, France 3 Tectospin, Institut Lavoisier de Versailles (UMR 8180), Universite´ de Versailles Saint Quentin en Yvelines, 45 Avenue des Etats-Unis, 78035 Versailles cedex, France 4 Institut Charles Gerhardt Montpellier, UMR 5253, CNRS-UM2-ENSCM-UM1, Physicochimie des Mate´riaux De´sordonne´s et Poreux, Universite´ de Montpellier II, Place Euge`ne Bataillon, C.C. 1503, 34095 Montpellier Cedex 5, France

5.1

Introduction

We show in this review that advanced solid state nuclear magnetic resonance (NMR) methods coupled with empirical and ab initio calculations of relevant parameters are well placed to meet the challenges provided by modern material chemistry. We demonstrate the

Functionalized Inorganic Fluorides: Synthesis, Characterization & Properties of Nanostructured Solids  2010 John Wiley & Sons, Ltd

Edited by Alain Tressaud

142

Functionalized Inorganic Fluorides

ability of these techniques on Al-based fluorides not only to characterize ordered crystalline structures but also to give detailed insights into the structure of partially disordered, amorphous or glassy materials. For these compounds, solid state NMR is a privileged technique since they contain at least two 100 % abundant and sensitive nuclei, 19F (nuclear spin I ¼ 1/2) and 27Al (I ¼ 5/2) which is a quadrupolar nucleus. This chapter is divided into three parts. In the first part (Section 5.2), various solid state NMR techniques, such as magic angle spinning (MAS), cross polarization (CP), multiple-quantum (MQ) and multidimensional spectroscopies, pertinent to obtain high resolution NMR spectra in Al-based fluorides, are presented. They allow the precise determination of the 19F isotropic chemical shift iso and 27 Al iso and quadrupolar parameters or reveal 19F-19F or 19F-27Al proximities. In the second part (Section 5.3), the strong potential of these techniques for the study of functionalized Al-based inorganic fluorides, synthesized for their possible application as catalysts (fluorinated alumina and zeolites, high surface area aluminium trifluorides and nanostructured aluminium fluoride hydrates and hydroxyfluorides), is shown. The compounds used as model for the assignment of the 19F and 27Al NMR lines in these materials are also presented. The third part (Section 5.4) is devoted to the determination of the 19F and 27Al NMR parameters of crystalline alkali and alkaline-earth fluoroaluminates. In the crystalline phases, the modelling of these parameters, starting from diffraction data and using semiempirical and Density Functional Theory (DFT) methods, aids spectral assignment. For 19F, double-quantum single-quantum (DQ-SQ) MAS NMR correlation experiments are demonstrated to be a complementary tool for correct line assignments.

5.2

High Resolution NMR Techniques

The quest for increased spectral resolution is a common feature of many solid-state NMR spectroscopists and has induced the developments of higher homogeneous magnetic fields (22 T spectrometers are now commercially available) and faster MAS (now up to 70 kHz) and of new NMR experiments, the ultimate goal being to obtain a spectrum on which all individual spectral signatures are resolved. 5.2.1

Fast MAS and High Magnetic Field

High resolution 1D 19F MAS NMR spectra can usually be recorded at moderate magnetic fields (7–10 T) under fast MAS (30–35 kHz), combined with heteronuclear decoupling when necessary. Nevertheless, because of the high gyromagnetic ratio and 100 % natural abundance of the 19F nucleus, 19F MAS NMR spectra may suffer from homogeneous line broadening due to residual homonuclear dipolar couplings not averaged out by fast MAS. This is shown in Figure 5.1 for b-CaAlF5 [1]; this compound contains five inequivalent F sites and only four resonances are resolved on the 19F MAS NMR spectrum recorded at 7 T and 35 kHz. The use of higher magnetic fields and/or ultra fast MAS (up to 70 kHz) allows this limitation to be overcome. In b-CaAlF5 [1,2], the five 19F resonances are resolved at

Short and Medium Range Order in Al-based Fluorides

35 kHz

* **

* **

* **

30 kHz

* **

* * *

25 kHz

* **

* *

20 kHz

* **

*

* * * **

15 kHz

* **

*

* * *

10 kHz 200

150

143

**

*

100

*

**

50

0

**

* *

* **

**

–50

* –100

* –150

Isotropic chemical shift (ppm) Figure 5.1 -CaAlF5 19F MAS NMR spectra collected at 7 T and at various spinning frequencies. The star symbols indicate the spinning sidebands [1]. Reprinted with permission from Inorg. Chem., 43, 2474–2485 (2004). Copyright 2004 American Chemical Society

17.6 T (Figure 5.2). For Na5Al3F14 [3] a dramatic improvement in resolution is also obtained at 19.6 T and very fast MAS (40 kHz). The increase in field enlarges the chemical shift difference (units: Hz) between resonances, reducing the spectral overlap between resonances. For quadrupolar nuclei, i.e. nuclei with spin > 1/2, an additional difficulty arises from the field-dependent shift and the line broadening by the second-order perturbation of quadrupolar interactions under the dominant Zeeman interaction [4]. MAS assists

144

Functionalized Inorganic Fluorides

17.6 T

*

7T

35

30

25

20

15

10

5

0

–5

–10

Isotropic chemical shift (ppm)

Figure 5.2 19F MAS NMR spectra of -CaAlF5 collected at 7 T for  r ¼ 35 kHz [1] and collected at 17.6 T for  r ¼ 34 kHz [2]. The symbol * indicates a nonidentified impurity

resolution but does not completely remove the anisotropic broadening if the quadrupolar interaction is large [5]. The second-order quadrupolar effect being inversely proportional to the square of the Larmor frequency, a quadratic gain both in resolution and sensitivity is expected at higher magnetic fields [6]. This is of major importance in terms of applicability of solid-state NMR to crystalline and amorphous inorganic materials containing quadrupolar nuclei such as 27Al [7]. This is shown in Figure 5.3 for nanostructured aluminium

7T 17.6 T

40

0

–40 δiso (ppm)

–80

–120

Figure 5.3 Central transition of the 27Al NMR spectra of the low surface area (LSA) AlF2.4(OH)0.6 sample recorded at 7 T and 17.6 T. The spinning frequency is 30 kHz for both [8]. Reprinted with permission from Chem. Mater., 20, 1459–1469 (2008), copyright 2008 American Chemical Society

Short and Medium Range Order in Al-based Fluorides

145

hydroxyfluorides derived from b-AlF3 [8]. The high magnetic field, by reducing the influence of the contribution of the quadrupolar interaction to the spectrum, allows the resolution of 27Al resonances with close iso values. As shown in Section 5.3 of this chapter, 27Al high-field NMR is a powerful tool to probe and quantify the various aluminium species in aluminium based fluorides.

5.2.2 5.2.2.1

27

Al NMR SATRAS Experiments

Satellite Transition Spectroscopy (SATRAS) [9,10] has been proven to be a useful technique to determine the NMR parameters of quadrupolar spin systems, in particular when the quadrupolar frequency  Q is so small that the structure of the central transition (CT) 1/2 ! 1/2 due to second-order effects is not resolved. In that case, reliable information can be obtained through the reconstruction of the spinning sideband manifold of the satellite transitions. The extent of the NMR spectrum gives  Q while the shape of the envelope of the spinning sidebands provides the asymmetry parameter Q. When the effects of pulse duration can be neglected (i.e. using short pulse and small flip angle), the SATRAS spectrum is quantitative. As an example, the 27Al NMR spectrum of -BaCaAlF7, which contains one aluminium site, is presented in Figure 5.4. The shapeless central transition and the overall SATRAS spectrum expansion indicate a low  Q value. The reconstruction of the SATRAS spectrum (Figure 5.4) was achieved using a homemade code based on the theoretical treatment developed by Skibsted et al. [9,11], and including a correction for the second-order frequency shift [12,13]. In this way, we precisely determined the 27Al NMR parameters (iso,  Q and Q) in numerous alkali and alkaline-earth fluoraluminates [2,14–17].

5.2.2.2

MQ-MAS Experiments

In 1995, Frydman and Harwood [18] introduced the multiple-quantum MAS (MQ-MAS) experiment, which makes use of the multi-quanta transitions of a quadrupolar nucleus. This two-dimensional (2D) NMR experiment correlates a high resolution isotropic spectrum (F1 vertical dimension) to the anisotropic central transition of the MAS spectrum (F2 horizontal dimension). The development of this method has considerably extended the applicability of solid-state NMR to quadrupolar nuclei in multisite crystalline phases and glasses. In the presence of strong heteronuclear dipolar couplings (such as 19F-27Al), resolution of the 27Al MQ-MAS spectrum can also be improved by applying composite decoupling schemes during the MQ evolution and the acquisition periods, as shown for Na5Al3F14 [19]. In crystalline compounds, the separation of the aluminium sites in two dimensions allows extraction of their quadrupolar interaction parameters individually [2,14–17,19] (Figure 5.5). In disordered crystalline compounds, an improved resolution of the resonances is obtained, as shown on the 27Al MQ-MAS spectrum of a high surface area (HS) pyrochlore aluminium hydroxyfluoride [20].

146

Functionalized Inorganic Fluorides

(b)

12

9

6

300

200

3

0

–3

–6

–9

–200

–300

–12

(a)

400

100 0 –100 frequency (kHz)

–400

Figure 5.4 Experimental (bottom) and reconstructed (top) spectra: (a) 27Al SATRAS NMR spectra [17] and (b) central transitions for -BaCaAlF7 collected at 7 T for  r ¼ 10 kHz. In (a) the central transitions are truncated to improve the visibility of the spinning sidebands. (a) Eur. J. Inorg. Chem., 1980–1988 (2007), Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission

5.2.2.3

TOP Spectra

Another way to increase resolution while avoiding an excessively long experimental time is to construct a two-dimensional one pulse (TOP) spectrum [21–23] from a conventional 1D MAS spectrum. The TOP spectrum evidences the well-known narrowing of the inner satellite transition line width: when the components are not well resolved on the N ¼ 0 cross-section, the projection over the satellite transitions (N 6¼ 0) shows a much better resolution as shown in HS pyrochlore aluminium hydroxyfluoride [20] (Figure 5.6), nanostructured Al-based fluoride-oxide materials [24] and aluminium fluoride hydrate with cationic vacancies [25].

Short and Medium Range Order in Al-based Fluorides (a)

147

(b)

Isotropic dimension (kHz)

0 2 2

0

–2 –4 –6 –8 Frequency (kHz)

–10

–12

0

–2 –4 –6 –8 Frequency (kHz)

–10

–12

4

(c) 6 8

2

0

–2 –4 –6 –8 MAS dimension (kHz)

–10

–12 2

Figure 5.5 (a) Experimental 27Al 3Q-MAS NMR spectrum of g-BaAlF5 collected at (a) 7 T and  r ¼ 25 kHz. The curves at the top and to the right of the spectrum are the full projections of the spectrum onto the MAS (F2) and isotropic (F1) dimension, respectively. The dashed lines indicate the F1 slices represented in (b) and (c). These slices are used for estimation of the NMR parameters. The reconstructed slices are shown below each experimental slice

3

(a)

4

(b)

2

1

N ≠ 0 ST

F1 (kHz)

12

16

1

20

2 3 4

24 20

10

0 –10 –20 –30 frequency (ppm)

27Al

–40

Spinning sideband order

N = 0 CT

–15 –10 –5 0 5 10 15 20

10

0 –10 –20 –30 frequency (ppm)

–40

27Al

Figure 5.6 (a) 3Q-MAS spectrum and (b) 2D TOP spectrum of HS pyrochlore aluminium hydroxyfluoride, AlF1.8(OH)1.2  0.3H2O. The N ¼ 0 cross-section shows the central transition spectrum and the N 6¼ 0 sum is the satellite transition spectrum showing enhanced resolution. The TOP spectrum was constructed by stacking subspectra shifted by the spinning frequency from a 1D MAS NMR spectra collected at  r ¼ 30 kHz and 17.6 T [20]. J. Mater. Chem., 18, 2483–2492 (2008). Reproduced by permission of the Royal Society of Chemistry

148

Functionalized Inorganic Fluorides

5.2.3

High Resolution Correlation NMR Techniques

An important part of the power of solid-state NMR methods comes from their ability to produce homonuclear or heteronuclear correlation charts that can evidence spatial proximity from dipolar interaction [26,27] or chemical bonding from J-couplings [26,28]. On the one hand, the J-based NMR experiments make use of the nonvanishing isotropic part of the indirect J-coupling. Characterizing the existence of a chemical bond, these NMR methods can provide very detailed insight into the structure of crystalline or glassy materials [26,28–32]. However, to date, 19F-27Al J-coupling has never been observed or used in the solid state. On the other hand, the dipolar-based experiments make use of the dipolar interaction, which vanishes under MAS. Reintroduction of this through-space interaction (which depends inversely upon the distance between the spins to the third power) can be conveniently achieved using various recoupling scheme (continuous wave or modulated radio frequency pulses in CP experiments). 19F-27Al or 19F-19F dipolar-based experiments have been widely used to characterize Al-based fluorides and will be presented in the following.

5.2.3.1

1D CP-MAS Experiments

CP is used almost ubiquitously in NMR of spin-1/2 nuclei, principally as a means of enhancing signal to noise [33–34]. CP involving a quadrupolar nucleus is a more complex process, the dynamics of which are still poorly understood [5]. Although signal enhancements are rarely observed for many quadrupolar nuclei, the application of CP to quadrupolar systems still holds great potential as a spectral editing tool for the elucidation of spatial relationships. At the simplest level, the detection of a CP signal indicates a spatial proximity between two heteronuclei. Several 19F-27Al CP-MAS experiment results will be described in Section 5.3.

5.2.3.2

1D Double Resonance NMR Experiments

Among the double resonance experiments, REDOR (rotational echo double resonance) [35], TRAPDOR (transfer population in double resonance) [36,37] and REAPDOR (rotational echo adiabatic passage double resonance) [38] were applied to probe the presence and strength of heteronuclear 19F-27Al dipolar couplings in Al-based fluorides. The dipolar coupling between the nuclear spin species S (whose signal is detected) and spatially close nuclear spin species I is reintroduced into the experiments by coherent I-spin irradiation during the rotor period. If heteronuclear dipoledipole interactions are present, the I-spin irradiation causes a decrease in S-spin signal intensity, relative to a reference experiment without I-spin irradiation (intensity S0) as shown in Figure 5.7 [39]. The magnitude (S0  S)/S0 of the difference signal depends on both the strength of the dipole-dipole coupling and on the length of the overall dipolar evolution time. In an S{I} REDOR experiment dipolar recoupling is accomplished by applying p-pulses on the I channel in the middle of the rotor period. Although REDOR has successfully been applied to systems where one or other of the dipolar-coupled nuclei is quadrupolar (S or I),

Short and Medium Range Order in Al-based Fluorides

149

F-Ca(3)

Al-F-Ca(n)

a

control

difference 0

–50 –100 –150 relative frequency (ppm)

–200

–250

Figure 5.7 19F{27Al}TRAPDOR spectra of crystalline Ca2AlF7. Note the large change in relative intensities of the two groups of central peaks (filled in), representing F-Ca(3) and Al-F-Ca(n) fluoride sites. The label ‘a’ marks the signal from a small amount of crystalline CaF2 impurity, which also disappears in the difference spectrum [39]. Reprinted from J. Non-Cryst. Solids, 337, 142–149 (2004), copyright 2004 Royal Society of Chemistry, with permission from Elsevier

as for CP, the spin dynamics associated with the experiment become much more complex than those for the simple two-spin-1/2 case, and the experiment itself may become somewhat inefficient [5]. The REAPDOR and TRAPDOR experiments were designed specifically to study dipolar interactions involving quadrupolar nuclei. These sequences are similar to REDOR in their philosophy, preventing the refocusing of the dipolar interaction under MAS, but S{I} TRAPDOR experiments utilize continuous-wave irradiation and S{I} REAPDOR adiabatic mixing pulses on the I (quadrupolar) spin. 19 F{27Al} and 27Al{19F} REDOR experiments and 19F{27Al} TRAPDOR and REAPDOR experiments provided information on spatial proximities between 19F and 27 Al in many Al-based fluorides. We can cite aluminium fluoride phosphate glasses [40,41], aluminosilicate glasses [42] and zeolites dealuminated with NH4SiF6 [43,44]. Other examples are detailed in Section 5.3. To be exhaustive about 1D double resonance NMR experiments, we have to mention the TEDOR (transferred echo double resonance) experiment, which involves transferring magnetization between two heteronuclear spins [45]. Nevertheless, this experiment has been applied only once to Al-based fluoride, on AlPO4-CJ2, a microporous

150

Functionalized Inorganic Fluorides

oxy-fluorinated aluminophosphate, by Amoureux et al., as a preliminary experiment for designing the TEDOR-MQMAS method [46]. 5.2.3.3

2D Double Resonance NMR Experiments

In many cases, the spectral resolution achieved merely by MAS is not sufficient to enable accurate determination of site-specific structural information. The MQMAS method described previously must therefore be employed to improve resolution and then combined with the spectral editing power of CP. In this way, the combination of MQMAS with CP – more precisely the initial method proposed and termed CP-MQMAS [47], is used for 19 F and 27Al spin pair. The efficiency of this method was tested on a fluorinated triclinic chabazite-like AlPO4 aluminophosphate (AlPO4-CHA), (AlPO4)3(C4H10NO)F, which contains three Al sites whose NMR parameters were previously determined [48] (Table 5.1). The CP-MQMAS method involves CP from I to the quadrupolar S spin single-quantum central-transition coherences. Then these are converted (by the use of a central-transition selective 90 pulse) into a population difference across the Zeeman mS ¼ – 1/2 eigenstates from which MQ coherence is then generated for the following MQMAS experiment. This ensures that MQ coherences are generated only for spins that have a dipolar coupling to I (i.e., an editing of the spectrum occurs). It is also possible to combine REDOR with MQMAS. Two approaches, MQ-t2-REDOR [49] and MQ-t1–REDOR [50], were presented and ascertained on AlPO4-CHA. An easier experiment, termed MQMAS with dipolar dephasing (DDMQMAS), was also proposed [51] and tested on the same compound. Finally, as previously mentioned, the combination of MQMAS and TEDOR is also possible (in an analogous manner to that of MQMAS and CP) as recently shown, by Amoureux et al., on AlPO4-CJ2 [46]. 5.2.3.4

2D SLF Experiments

To monitor the effective aluminium-fluorine dipolar couplings in cryolite, as a function of temperature, amplified 2D separated local field (SLF) NMR experiments under the action of fast MAS (which is in turn desirable for the simple elimination of the homonuclear 19 19 F- F couplings) were recorded by Kotecha et al. [52]. In such SLF MAS experiments rotor-synchronized pulses were applied to achieve a net heteronuclear dipolar evolution with variable amplification factors xN of the 27Al-19F interaction along the indirect domain (x2 SLF, x4 SLF, and x8 SLF), followed by observation of aluminium’s central-transition {19F}-decoupled evolution along the direct domain [53]. 5.2.3.5

2D MAS CP-HETCOR Experiments

Heteronuclear correlation (HETCOR) solid-state NMR spectroscopy [54] has been widely used to provide information on the spatial proximity of different nuclei in complex spin systems. When more than one distinct I and/or S spin species is present, a 2D correlation experiment enables the detection of dipolar couplings between specific distinct I–S pairs. Typically, the experiment consists of a 90 pulse that creates transverse I spin magnetization, which evolves for a time t1 before it is transferred to spin S, usually via CP. The S spin FID is then detected in t2. 2D Fourier transform yields a 2D spectrum, with the appearance of cross peaks between individual I and S resonances from spins which are dipolar coupled.

Short and Medium Range Order in Al-based Fluorides

151

These 2D HETCOR experiments are also feasible when either I or S is a half-integer quadrupolar nucleus [55,56]. The spectra are able to provide information unavailable by 1D methods due to the increase in resolution obtained from the presence of a second spectral dimension. They are particularly informative if at least one of the two dimensions of the 2D spectrum is well resolved as shown on b-BaAlF5 (Figure 5.8) [57]. 19F-27Al MAS CP-HETCOR experiment was also applied in zeolite [44] and chiolite [58]. Other results will be described in Section 5.3.

(b)

(a) Al proj. Al1

19

1 cross section Al1

2

–140

3

4 5

cross section Al2

6

8

–120

7 8

F projection

9

10

9

–100

5

7

4

6

–120

–140

3 2

1

–160

19

–100

F SQ dimension (ppm)

–160

Al2

10

10

0 27

F SQ dimension (ppm)

–10 –20 –30 –40 –50 –60 Al SQ dimension (ppm)

Figure 5.8 (a) 2D 19F-27Al MAS (25 kHz) CP-HETCOR correlation spectrum of -BaAlF5. Top spectra are the full projection onto the 27Al dimension and simulation of Al1 and Al2 resonances. Dash lines indicate position where the cross-sections of Al1 and Al2 were extracted from. Right spectrum is the full projection in the 19F dimension, on which lines are labelled. (b) Selected cross-section of Al1 and Al2 and full projection in the 19F dimension of the CPHETCOR spectrum. 19F lines are labelled. Dash lines indicate 27Al-19F correlations [57]. Phys. Chem. Chem. Phys., 11, 950–957 (2009). Reproduced by permission of the PCCP Owner Societies Copyright (2009) Royal Society of Chemistry

5.2.3.6

19

F DQ-SQ MAS Correlation Experiments

Homonuclear fluorine spatial proximities can be evidenced through 2D 19F DQ-SQMAS correlation NMR experiments [59] as shown in oxyfluoride [60] and in fluoride materials [57,61–63] leading to the assignment of 19F MAS NMR spectra. In these examples, the 19 19 F- F homonuclear interaction, which is averaged out by fast MAS, is reintroduced using the back-to-back (BABA) recoupling sequence [64]. In a 2D 19F MAS DQ-SQ NMR spectrum, fluorine atomic proximities between inequivalent F sites are revealed by paired cross-correlation peaks appearing at the individual chemical shifts of the two dipolarcoupled nuclei in the SQ direct dimension and the sum of them in the indirect DQ dimension, while the proximity between two equivalent F sites is disclosed by a single auto-correlation peak located on the DQ diagonal (with a slope of 2) of the 2D spectrum

152

Functionalized Inorganic Fluorides

(Figure 5.9). Assuming that the 19F multi-spin homonuclear dipolar interaction is averaged out by the fast MAS frequency (such that the recoupled spin system can be considered as an ensemble of simple spin pairs) and using short DQ excitation/reconversion periods, the intensity of the cross-peaks in the DQ-SQ rotor-synchronized spectrum is then expected to be proportional to the number of spin pairs and to D2t2, with D the dipolar coupling constant and t the recoupling time [57 and references therein]. Advantages and limitations of this dipolar-based 2D NMR experiments for 19F MAS spectra assignments of fluoroaluminates were recently presented [57] and will be discussed in Section 5.4.1.3.

9

10

8

7

5

6

4

3

1

2

10-9 2-1

–300

3-1 4-1

5-1 7-4

6-2

8-2 8-3 8-4

5-5

5-2 4-3

6-1

3-2

7-1

–275

7-2 7-3 9-1

8-5

9-2

6-5

9-3

–250

19F

DQ dimension (ppm)

8-1

7-6

9-4 9-5

10-2

6-4

6-6

8-7

10-3 10-4

10-5 9-8

10-6 10-7

–225

10-8

–90

–100

–110 19F

–120 –130 SQ dimension (ppm)

–140

–150

Figure 5.9 2D 19F DQ-SQ MAS (30 kHz) NMR correlation spectrum of -BaAlF5. Top spectrum is the full projection in the 19F SQ dimension, on which lines are labelled. The DQ diagonals (with a slope of 2) of the 2D spectrum on which autocorrelation peaks appear are indicated by the dash lines. Paired cross-correlation peaks are indicated by horizontal solid lines [57]. Phys. Chem. Chem. Phys., 11, 950–957 (2009). Reproduced by permission of the PCCP Owner Societies Copyright (2009) Royal Society of Chemistry

Short and Medium Range Order in Al-based Fluorides

5.3

153

Application to Functionalized Al-Based Fluorides with Catalytic Properties

In this section we report on structural information obtained by high resolution solid state NMR on various aluminium fluoride phases and fluorinated aluminas and zeolites, which are extensively studied for their catalytic potential. The change in the coordination environment of aluminium on fluorination and the high electronegativity of fluorine atom are reported to be responsible for the observed modification of the acidity, and thus catalytic activity, of these materials [65–67]. Then, identification of the bulk and surface species is critical for the understanding of the catalytic properties. In the studies discussed in the following, assignment of 19F and 27Al NMR resonances of fluorinated aluminas and zeolites was achieved on the basis of model compounds such as AlF3 and AlF3  3H2O [68–69].

5.3.1

Crystalline Aluminium Fluoride Phases

AlF3 forms a rich variety of phases. All their structures are based on corner connected AlF6 octahedra. Two of them were studied by solid state NMR: -AlF3 (27Al [14,68], 19F [68,70]) and b-AlF3 (27Al and 19F [68]). Table 5.1 gathers structural information and 19F and 27Al NMR parameters. For these two polymorphs, NMR studies provide 19F iso values typical of fluorine atoms bridging two aluminium atoms (BF) and 27Al iso values typical of AlF6 octahedra. Two metastable phases of aluminium trifluoride (b- and -AlF3) and milled -AlF3 are known as suitable catalysts [71–73]. Crystalline AlF3 phases in general suffer from a low surface area (